JP3824749B2 - Bone image processing method and apparatus - Google Patents

Description

【００２５】
であり、定義域内で値が０で、その定義域Ｇが下記式（４）であるとする。
【００２６】
【数４】

【００２７】
このとき、モーフォロジー演算の基本形は式（５）〜（８）に示すように、非常に簡単な演算となる。
【００２８】
【数５】

【００２９】
すなわち、ダイレーション（dilation）処理は、注目画素を中心とした、±ｍ（構造要素Ｂに応じて決定される値であって、図10中のマスクサイズに相当）の幅の範囲内の最大値を探索する処理であり（同図（Ａ）参照）、一方、エロージョン（erosion ）処理は、注目画素を中心とした、±ｍの幅の範囲内の最小値を探索する処理である（同図（Ｂ）参照）。また、オープニング（opening ）処理はエロージョン処理後にダイレーション処理を行なう処理、すなわち最小値の探索の後に最大値を探索する処理であり、クロージング（closing ）処理は、ダイレーション処理後にエロージョン処理を行なう処理、すなわち最大値の探索の後に最小値を探索する処理に相当する。
【００３０】
つまりオープニング処理は、低輝度側から濃度曲線ｆ（ｘ）を滑らかにし、マスクサイズ２ｍより空間的に狭い範囲で変動する凸状の濃度変動部分（周囲部分よりも輝度が高い部分）を取り除くことに相当する（同図（Ｃ）参照）。
【００３１】
一方、クロージング処理は、高輝度側から濃度曲線ｆ（ｘ）を滑らかにし、マスクサイズ２ｍより空間的に狭い範囲で変動する凹状の濃度変動部分（周囲部分よりも輝度が低い部分）を取り除くことに相当する（同図（Ｄ）参照）。
【００３２】
なお、構造要素ｇが原点に対して対称ではない場合の、式（５）に示すダイレーション演算をミンコフスキー（Minkowski ）和、式（６）に示すエロージョン演算をミンコフスキー差という。
【００３３】
ここで、濃度の高いもの程大きな値となる高濃度高信号レベルの信号の場合においては、濃度値ｆ（ｘ）の画像信号値が高輝度高信号レベルの場合に対して大小関係が逆転するため、高濃度高信号レベルの信号に対するダイレーション処理と高輝度高信号レベルに対するエロージョン処理（同図（Ｂ））とは一致し、高濃度高信号レベルの信号に対するエロージョン処理と高輝度高信号レベルに対するダイレーション処理（同図（Ａ））とは一致し、高濃度高信号レベルの信号に対するオープニング処理と高輝度高信号レベルに対するクロージング処理（同図（Ｄ））とは一致し、高濃度高信号レベルの信号に対するクロージング処理と高輝度高信号レベルに対するオープニング処理（同図（Ｃ））とは一致する。
【００３４】
なお、本項では高輝度高信号レベルの画像信号（輝度値）の場合について説明する。
【００３５】
（石灰化陰影検出への応用）
石灰化陰影の検出には、原画像から平滑化した画像を引き去る差分法が考えられる。単純な平滑化法では石灰化陰影と細長い形状の非石灰化陰影（乳腺、血管および乳腺支持組織等）との識別が困難であるため、東京農工大の小畑らは、多重構造要素を用いたオープニング処理に基づく下記式（９）で表されるモーフォロジーフィルターを提案している（「多重構造要素を用いたモルフォロジーフィルタによる微小石灰化像の抽出」電子情報通信学会論文誌 D-II Vol.J75-D-II No.7 P1170 〜1176 1992年７月、「モルフォロジーの基礎とそのマンモグラム処理への応用」MEDICAL IMAGING TECHNOLOGY Vol.12 No.1 January 1994 等）。
【００３６】
【数６】

【００３７】
ここでＢi （ｉ＝１，２，…，ｎ）は、直線状の大きさがｍ画素でｎ個（例えば図11に示すものでは、９画素４方向でありｍ＝９，ｎ＝４）の構造要素（これらを全体として以下、ｍ画素ｎ方向の多重構造要素という）である。構造要素Ｂi を検出対象である石灰化陰影よりも大きく設定すれば、上記オープニング演算による処理で、構造要素Ｂi よりも細かな信号変化部分（空間的に狭い範囲で信号が変動する画像部分）であって周囲よりも輝度値の大きい凸状の部分である石灰化陰影は取り除かれる。一方、細長い形状の乳腺の陰影等の非石灰化陰影はその長さが構造要素Ｂi よりも長く、その傾きが４つの構造要素Ｂi のいずれかに一致すればオープニング処理（式（９）の第２項の演算）をしてもそのまま残る。したがってオープニング処理によって得られた平滑化画像（石灰化陰影のみが取り除かれた画像）を原画像ｆから引き去ることで、小さな石灰化陰影のみが含まれる画像が得られる。これが式（９）の考え方である。
【００３８】
なお、前述したように、高濃度高信号レベルの信号の場合においては、石灰化陰影は周囲の画像部分よりも濃度値が低くなり、石灰化陰影は周囲部分に対して濃度値の小さい凹状の信号変化部分となるため、オープニング処理に代えてクロージング処理を適用し、式（９）に代えて式（10）を適用する。
【００３９】
【数７】

【００４０】
なお、モーフォロジー演算の一例である式（10）のクロージング処理を具体的に説明する。
【００４１】
すなわち、高濃度高信号レベルの画像信号である濃度値Ｓorg についてのモーフォロジー演算によれば、例えば図12（１）の実線に示すような濃度値Ｓorg の分布を有する画像データに対して、同図（２）に示すような直線状の３画素の構造要素Ｂで、最大値処理（ダイレーション処理）を行うことにより、ある注目画素の濃度値Ｓ_i は、その注目画素を中心として互いに隣接する３画素（構造要素Ｂにより決定される）の中の最大値Ｓ_i+1 を採用したＳ_i ′に変換される。この演算を全画素について行うことにより、濃度値Ｓorg ′の分布を有する同図（１）の破線で示す最大値信号に変換される。
【００４２】
次に、この最大値処理により得られた最大値信号に対してさらに構造要素Ｂによる最小値処理（エロージョン処理）を考えると、同図（１）の破線で示された注目画素の最大値信号Ｓ_i ′は、その注目画素を中心として互いに隣接する３画素の中の最小値Ｓ_i-1 ′を採用したＳ_i ″（＝Ｓ_i ）に変換される。この演算を全画素について行うことにより、最大値処理後の最小値信号Ｓorg ″の分布は同図（１）の一点鎖線で示すものとされる。この一点鎖線で示された画像信号は、もとの実線のオリジナルの画像データに対して、構造要素Ｂよりも空間的に狭い範囲で信号が変動する画像部分が消え、構造要素Ｂよりも空間的に広い範囲で変動する信号値の変化部分である画像部分や変動のない画像部分はもとの形状のまま残っていることを示している。すなわち、以上の処理（クロージング処理）は、画像濃度の分布を高濃度側から平滑化する処理として作用する。
【００４３】
このようにクロージング処理で得られた値（Ｓorg に対して最大値処理を行なった後にさらに最小値処理を行なった値）を原画像信号Ｓorg から差し引くことにより得られた値Ｓmor は、上記クロージング処理で消された空間的に狭い範囲で変動する信号値の変化部分である画像部分を表す。
【００４４】
ここで、本来、画像信号は２次元の要素である位置（ｘ，ｙ）と、３次元目の要素である信号値ｆ（ｘ，ｙ）を有するが、上記説明においては、理解の容易化のために、この２次元上に展開された画像の所定の断面に現れた、１次元状の画像信号分布曲線について説明した。
【００４５】
したがって実際には、以上の説明を２次元画像に適用する必要があり、多重構造要素を用いるのも２次元画像に対応させるためである。
【００４６】
次にモーフォロジー演算に基づくスケルトン処理について説明する。
【００４７】
スケルトン処理とは一般に図形の骨格（skeleton）を抽出する処理であり、骨格とは図形に内接する円盤の中心の集合としてとらえることができる。すなわち例えば図13（Ａ）〜（Ｅ）に示す各図形（中太の実線で表記）の骨格はそれぞれ太い実線で示すものとなる。
【００４８】
以下、このスケルトン処理を、上述したモーフォロジー演算により行なう場合について説明する。この場合、スケルトン処理は下記式（１）または（２）で表すことができる。
【００４９】
【数１】

【００５０】
【数２】

【００５１】
ここで式（１）と式（２）とは前述した、画像を高濃度高信号レベルの画像信号として表すか、または高輝度高信号レベルの画像信号として表すかの差異によるものであり、高濃度高信号レベルの画像信号で表された画像から低濃度（高輝度）の画像部分の骨格を抽出する場合は式（１）を適用し、一方、高輝度高信号レベルの画像信号で表された画像から低輝度（高濃度）の画像部分の骨格を抽出する場合は式（２）を適用するものであり、作用自体の実質的な差はない。
【００５２】
例えばネガフイルム（高濃度高信号レベル）上においては骨部は他の画像部分に比べてその濃度は低いものとなり、骨梁の存在する部分は濃度が低く、存在しない部分は濃度が高くなる。したがって、周囲よりも濃度の低い部分となる骨梁を対象としてスケルトン処理を行なうことに該当するため式（１）を適用すればよい。
【００５３】
ここで式（１）における構造要素Ｂを半径ｒの円とし、図14に示す図形に対してスケルトン処理を施した状態を示す。図14に示す図形は、その輪郭よりも外側領域が濃度の高い部分であり、内側が濃度の低い部分である。
【００５４】
この図形に対してまず図示上段では構造要素Ｂによるエロージョン処理が施される。λ＝０（構造要素Ｂによる０回のエロージョン処理）では図形は何ら変化がない。
【００５５】
λ＝１（構造要素Ｂによる１回のエロージョン処理）では図形が構造要素Ｂの半径ｒ分だけ内側に埋め込められる。
【００５６】
λ＝２（構造要素Ｂによる２回のエロージョン処理）では図形の円から突出した部分が完全に消失する。
【００５７】
同様の操作を繰り返すことにより、λ＝Ｎ−１（構造要素ＢによるＮ−１回のエロージョン処理）で図形は半径ｒ以下の円のみとなる。
【００５８】
一方、図示中段は、この構造要素Ｂによる各回（λ＝０，１，２，…，Ｎ−１，Ｎ）のエロージョン処理を施した画像に対して、さらにそれぞれ構造要素Ｂによるオープニング処理を施した図形である。
【００５９】
この図示上段の図形から図示中段の図形を、処理の回数を対応させて差し引いた図形が図示下段である。
【００６０】
λ＝１において、元の図形の円から突出した部分の骨格要素が抽出され、λ＝Ｎ−１において、元の図形の円の骨格要素が抽出されていることが分かる。
【００６１】
このように元の図形に対してエロージョン処理を施し、さらにオープニング処理を施し、処理回数を対応させて差し引き、この結果の和集合を求めたものが式（１）の意味するところである。
【００６２】
式（２）では濃度の高低が式（１）とは逆転した図形から骨格要素を抽出するのに有効であり、元の図形に対してダイレーション処理を施し、さらにクロージング処理を施し、処理回数を対応させて差し引き、この結果の和集合を求めることにより骨格要素を抽出することを意味する。
【００６３】
なお上記式（１）、（２）中の和集合として、ｎ1 の値が比較的大きいもののみの和集合、すなわち例えばｎ1 ＝０，１の場合を除いたλ＝２，３，４，５等の和集合を採用し、この場合に得られる骨格要素を表示することにより、骨梁の変化をより見易くすることができ、好ましい。これは、λが０や１のときは特に、非常に空間周波数の高いノイズ成分も抽出されるため、和集合からこれらのノイズ成分を除外することによって、より読影性能の高い画像を得ることができるからである。以下の発明においても同様である。
【００６４】
本発明の第２の骨部画像処理方法は、軟部組織と骨部組織とを含む被検体を被写体とする、エネルギーサブトラクション処理用の放射線エネルギー分布が互いに異なる２つの放射線画像に基づいて、前記骨部組織における骨成分の密度に関する情報を取得する骨部画像処理方法であって、
前記２つの放射線画像に対してエネルギーサブトラクション処理を施すことにより、前記被写体中の骨部組織の画像が抽出又は強調された骨部画像を生成し、該骨部画像に基づいて、前記骨成分の密度に関する数値情報を取得するとともに、
前記２つの放射線画像のうち少なくとも一方、または該２つの放射線画像を画素を対応させて重み付け加算して得られた加算画像に対して、骨梁の強調処理を施すことにより該骨梁の形態情報を取得することを特徴とするものである。
【００６５】
すなわち、本発明の第１の骨部画像処理方法が、エネルギーサブトラクション処理のための２つの放射線画像をエネルギーサブトラクション処理した後の骨部画像に基づいて骨梁の強調処理を行なうものであるのに対し、本発明の第２の骨部画像処理方法は、エネルギーサブトラクション処理を行なうのに先だってエネルギーサブトラクション処理のための２つの放射線画像のうち少なくとも一方、または２つの放射線画像を画素を対応させて重み付け加算して得られた加算画像に対して、骨梁の強調処理を行なうものである。
【００６６】
また、本発明の第２の骨部画像処理方法においても、第１の骨部画像処理方法と同様に、前記骨梁の形態情報に基づいて骨梁構造を表す指標値を算出してもよく、さらにはこの骨梁構造を表す指標値および前記骨成分の密度に関する数値情報に基づいて、被検体の骨部組織の状態を判定するようにしてもよい。
【００６７】
なお、上記各発明において骨成分の密度に関する数値情報の取得方法としては、前記被検体のほか、その骨塩量が既知の骨塩レファレンス、すなわち人骨を模したレファレンスであってその骨塩量を段階的に変化させて作成されたもの（各段階における骨塩量は既知）を被写体に加えて放射線画像を生成し、骨部画像中に現れた骨塩レファレンスの画像の濃度と前記既知の骨塩量とから濃度・骨塩量変換テーブルを作成し、濃度・骨塩量変換テーブルに基づいて被写体の骨部画像における骨部の濃度を骨塩量に変換することにより取得する方法（例えば特開平４−11473 号等参照）などを適用することができる。
【００６８】
本発明の骨部画像処理装置は本発明の骨部画像処理方法を実施するための装置であって、軟部組織と骨部組織とを含む被検体を被写体とする、エネルギーサブトラクション処理用の放射線エネルギー分布が互いに異なる２つ以上の放射線画像に基づいて、前記骨部組織における骨成分の密度に関する情報を取得する骨部画像処理装置であって、
所定の２つの画像信号に基づいて、該画像信号が表す画像のうち所望とする骨部を抽出又は強調処理するエネルギーサブトラクション処理を施すエネルギーサブトラクション処理手段と、
該骨部の画像信号に基づいて、該骨成分の密度に関する数値情報を求める骨密度解析手段と、
前記エネルギーサブトラクション処理前の前記２つの放射線画像を表す各画像信号、前記エネルギーサブトラクション処理後の前記骨部の画像信号、または前記２つの放射線画像を表す各画像信号を重み付け加算して得られた加算画像信号に対して、骨梁を強調処理する骨梁強調手段とを備えたことを特徴とするものである。
【００６９】
また、本発明の骨部画像処理装置においては、前記骨梁強調手段により強調処理された骨梁の形態情報に基づいて、骨梁構造を表す指標値を算出する指標値算出手段を備えるものとしてもよく、さらに、この骨梁構造を表す指標値および前記骨成分の密度に関する数値情報に基づいて、被検体の骨部組織の状態を判定する判定手段を備えるものとしてもよい。
【００７０】
ここで上記骨梁強調手段としては式（１）又は（２）で示されるモーフォロジー演算に基づくスケルトン処理を施すモーフォロジー演算手段を適用することができる。
【００７１】
なお、モーフォロジー演算を適用する画像信号が高濃度高信号レベルの画像信号で表されるものである場合は、式（１）を適用し、モーフォロジー演算を適用する画像信号が高輝度高信号レベルの画像信号で表されるものである場合は、式（２）を適用する。
【００７２】
高濃度高信号レベルの画像信号で表される画像においては、骨梁は低濃度の画像部分として表現され、高輝度高信号レベルの画像信号で表される画像においては、骨梁は高輝度の画像部分として表現されるからである。
【００７３】
なお上記式（１）、（２）中の和集合として、ｎ1 の値が比較的大きいもののみの和集合、すなわち例えばｎ1 ＝０，１の場合を除いたλ＝２，３，４，５等の和集合を採用し、この場合に得られる骨格要素を表示することにより、骨梁の変化をより見易くすることができ、好ましい。
【００７４】
また前記骨密度解析手段としては、被検体とともに放射線画像に記録された、その骨塩量が既知の骨塩レファレンスの、前記骨部画像中に現れた骨塩レファレンスの画像の濃度と前記既知の骨塩量とから作成される濃度・骨塩量変換テーブルに基づいて、被写体の前記骨部画像における骨部の濃度を骨塩量に変換して、前記骨成分の密度に関する数値情報を求めるものを適用することができる。
【００７５】
【発明の効果】
本発明の骨部画像処理方法および装置は、放射線画像に対してエネルギーサブトラクション処理を施して骨部画像を得、この骨部画像に基づいて骨成分の密度に関する数値情報を取得するとともに、放射線画像または骨部画像に対して骨梁の強調処理を施すことにより骨梁の形態情報を取得することによって、骨梁の状態を視覚的にも認識しやすいものとし、特に骨粗鬆症の診断に有用な情報を提供することができる。
【００７６】
とくに、骨梁の形態情報に基づいて骨梁構造を表す指標値を算出することにより、骨梁の状態を数値として得ることができ、これにより骨粗鬆症の診断に有用な情報を定量的に求めることができる。
【００７７】
【発明の実施の形態】
以下、本発明の骨部画像処理装置の具体的な実施の形態について図面を用いて説明する。
【００７８】
図１は本発明の骨部画像処理装置の一実施形態を示すブロック図、図２は図１に示す骨部画像処理装置の作用の流れを示す図である。
【００７９】
図示の骨部画像処理装置100 は、軟部組織（A2，B2）と骨部組織（A1，B1）とを含む被検体を被写体とする、エネルギーサブトラクション処理用の高エネルギー画像Ａ（図３参照）と低エネルギー画像Ｂとに基づいて、骨部組織C1が強調された骨部画像Ｃを生成するエネルギーサブトラクション処理手段10と、この骨部画像Ｃを表す骨部画像信号（エネルギーサブトラクション画像信号）Ｓsub に基づいて、骨部組織C1における骨成分の密度に関する数値情報を求める骨密度解析手段30と、骨部画像信号Ｓsub に対して骨梁を強調処理する骨梁強調手段50とを備えた構成である。
【００８０】
ここで上記高エネルギー画像Ａと低エネルギー画像Ｂは、いわゆる１ショットエネルギーサブトラクション（特開昭59− 83486号等参照）により１回の撮影で得られたエネルギー分布が互いに異なる画像であり、各画像にはそれぞれ被検体の画像とともに、被検体とともに撮影された骨塩レファレンス31（図４参照）の画像A3，B3が記録されている。
【００８１】
骨塩レファレンス31は、骨塩の量すなわちCaCO₃ の含有量（wt％）を段階的に変えてそれを異なるセクションd1，d2，d3，d4，d5，d6に並べた構造であり、すべてのセクションにおけるCaCO₃ の含有量はそれぞれ既知である。
【００８２】
またエネルギーサブトラクション処理手段10は、入力された高エネルギー画像Ａを表す高エネルギー画像信号Ｓａと、低エネルギー画像Ｂを表す低エネルギー画像信号Ｓｂとに基づいて、低エネルギー画像Ｂの軟部組織B2と高エネルギー画像Ａの軟部組織A2との濃度が一致するように、両画像に重み付けをしたうえで両信号の差（差信号＝エネルギーサブトラクション画像信号）Ｓsub を演算する。この結果、エネルギーサブトラクション画像信号Ｓsub は、骨部組織C1が強調された骨部画像を表すものとなる。
【００８３】
骨密度解析手段30は、上記骨部画像Ｃ中に現れた骨塩レファレンス31の画像C3の、各セクションd1，d2，d3，d4，d5，d6ごとの濃度を求め、この各セクションd1，d2，d3，d4，d5，d6ごとの濃度と各セクションd1，d2，d3，d4，d5，d6ごとの既知の骨塩量とを対応させて、図５のグラフに示すような対応関係の濃度・骨塩量変換テーブルを作成する。そして骨部画像Ｃ中に現れた骨部組織C1の濃度を求め、この求めた濃度を作成した濃度・骨塩量変換テーブルにしたがって骨塩量に変換することで、骨成分の密度に関する数値情報を求めるものである。
【００８４】
骨梁強調手段50は、下記式（１）で示されるモーフォロジー演算に基づくスケルトン処理を施すことによって、骨部画像Ｃにおける骨部組織C1の骨梁を強調するものである。
【００８５】
【数１】

【００８６】
次に本実施形態の骨部画像処理装置100 の作用について説明する。
【００８７】
エネルギーサブトラクション処理手段10は入力された高エネルギー画像Ａを表す高エネルギー画像信号Ｓａと、低エネルギー画像Ｂを表す低エネルギー画像信号Ｓｂとから、下記式（11）に基づくエネルギーサブトラクション処理を施して骨部画像信号Ｓsub を求める。
【００８８】
Ｓsub ＝Ｋ logＳａ−Ｌ logＳｂ＋Ｍ （11）
（但し、Ｋ、Ｌ、Ｍは係数を表わす）
得られた骨部画像信号Ｓsub は、軟部組織A2，B2が消去された、相対的に骨部組織A1，B1が強調された骨部画像Ｃを表すものとなる。
【００８９】
次いで、このようにして得られた骨部画像Ｃを表す骨部画像信号Ｓsub は骨密度解析手段30に入力され、骨密度解析手段30は、骨部画像Ｃ中に現れた骨塩レファレンス31の画像C3の、各セクションd1，d2，d3，d4，d5，d6ごとの濃度（画像信号）を求める。
【００９０】
そして各セクションd1，d2，d3，d4，d5，d6ごとの濃度と各セクションd1，d2，d3，d4，d5，d6ごとの既知の骨塩量（CaCO₃ 量）とを対応させて、濃度・骨塩量変換テーブル（図５参照）を作成する。その後、骨部画像信号Ｓsub を濃度・骨塩量変換テーブルにしたがって骨塩量に変換することによって、骨部組織C1における骨成分の密度に関する数値情報を求める。
【００９１】
得られた数値情報は表示装置等の出力媒体に出力され、診断に供される。
【００９２】
一方、骨部画像Ｃを表す骨部画像信号Ｓsub は骨梁強調手段50にも入力される。
【００９３】
骨梁強調手段50は、入力された骨部画像信号Ｓsub に対して、前記式（１）で示されるモーフォロジー演算に基づくスケルトン処理を施し、骨部画像Ｃにおける骨部組織C1の骨梁を強調処理する。なお、このモーフォロジー演算に用いられる構造要素は、骨梁の骨格を抽出するのに適した大きさ（具体的には骨梁の太さに応じた大きさ）のものを設定する。
【００９４】
骨梁は骨の内部に縦横に延びる海綿質であり、放射線画像においては図６に示すように、この縦横に延びる海綿質が幾重にも重なって写し出されるが、骨梁強調手段50によるモーフォロジー演算に基づくスケルトン処理によって、図７に示すように海綿質の骨格だけが幾重にも重なった画像とされる。
【００９５】
そして、この骨梁の骨格だけが抽出された画像は外部の表示手段等に出力される。
【００９６】
このように、本実施形態の骨部画像処理装置によれば、骨成分の密度を数値情報として定量的に得ることができるとともに、その骨梁の状態を視覚的な画像情報としても得ることができるため、骨梁の状態の診断に非常に有効であり、特に骨粗鬆症の診断に有用な情報を提供することができる。
【００９７】
なお、図１に示した実施形態の骨部画像処理装置は、骨梁強調手段50において骨梁の画像情報を取得するものであるが、図15のフローチャートに示すように、骨梁強調処理手段50において得られた骨梁の画像情報に基づいて、骨梁構造を示す指標値を算出し、この指標値と骨密度の数値情報とを判定装置に入力して、被検体の骨部組織の状態を判定するするようにしてもよい。以下、この指標値の算出について説明する。
【００９８】
骨梁構造の指標値を算出する方法の具体例としては、Star volume 法が挙げられる。このStar volume 法は、図16に示すように、骨髄腔内のある点から全ての方向について骨梁に遮られることなく見渡せる範囲の骨髄腔体積の平均を意味するmarrow space star volume（Ｖｍ）および同様に骨梁の中の点から全ての方向に骨梁の端に至る範囲の体積の平均を意味するtrabecular star volume（Ｖｔ）の２種類が定義されている。なお、図16においては、斜線部が骨梁を、その他の部分が骨髄腔を示す。Star volume は、標本のサンプリング法の工夫により骨髄腔の骨梁の大きさをｍｍ³ あるいはμｍ³ といった三次元の値として表す隔たりのない立体学的指標とされている。Ｖｍは骨梁の連続性が高い場合は小さく、骨梁の消滅や穿孔が多いときは大きくなり、逆にＶｔは骨梁の連続性が高い場合は大きく、骨梁の消滅や穿孔が多いときは小さくなる。
【００９９】
骨髄腔内の任意の点ｉにおけるＶｍｉは以下の式(12)により定義される。
【０１００】
Ｖｍｉ＝π／３×ｌ_0m ³ (12)
但し、ｌ₀ は点ｉを中心とする任意の方向において、骨髄腔が連続する長さを示す。また、ｌ_0m ³ は点ｉを中心として放射状に全ての方向においてｌ₀ ³を求めた場合の平均値を表す。
【０１０１】
次に、骨梁上の任意の点ｊにおけるＶｔｊは以下の式(13)により定義される。
【０１０２】
Ｖｔｊ＝π／３×Σｌ₁ ⁴／Σｌ₁ (13)
但し、ｌ₁ は点ｊを中心とする任意の方向において、骨梁が連続する長さを示す。また、ｌ₁ は点ｊを中心として放射状に全ての方向において求められ、Σはその全ての方向における和を求めることを表す。そして、本実施形態においては、ＶｍｉおよびＶｔｊをサンプリング点ごとに算出し、その平均値をＶｍおよびＶｔとした。
【０１０３】
なお、骨梁構造の指標値としては、ＶｍとＶｔのいずれを用いてもよいが、本実施形態においてはＶｔを用いるものとする。
【０１０４】
そして、骨梁の画像情報を得た後、上記式(13)に基づいてＶｔの値を求め、Ｖｔの値を骨密度の数値情報とともに判定装置に入力する。
【０１０５】
判定装置においては、例えば、図17に示すように、被検体の年齢、骨密度値およびＶｔを軸とする三次元座標が形成され、入力された結果がこの三次元座標内のいずれの位置に存在するかに応じて、骨部組織の状態を判定するものである。すなわち、判定装置に入力された被検体の年齢、骨密度値およびＶｔが点Ｐ１で示す位置にあれば状態Ａにあると判定し、点Ｐ２で示す位置にあれば状態Ｂにあると判定し、点Ｐ３で示す位置にあれば状態Ｃにあると判定するものである。
【０１０６】
このように、骨梁の画像情報に基づいて骨梁の指標値を得ることにより、骨成分の密度とともに骨梁の状態をも数値情報として定量的に得ることができるため、骨梁の状態の診断に非常に有効であり、特に骨粗鬆症の診断に有用な情報を提供することができるとともに、被検体の骨部組織の状態を自動的に判定することができる。
【０１０７】
なお、図15に示す実施形態においては、図17に示すように、被検体の年齢、骨密度値およびＶｔを軸とする三次元座標を形成して被検体の骨部組織の状態を判定しているが、図18（ａ）に示すように、被検体の年齢と骨密度値とに基づいて仮の判定１〜３を求め、図18（ｂ）に示すように、判定１〜３とＶｔの値との関係を表すテーブルを参照して被検体の骨部組織の状態を判定するようにしてもよい。
【０１０８】
また、図15に示す実施形態においては、Ｖｔを指標値として求めているが、上記Ｖｍを指標値として求めてもよく、さらには、指標値をNode-strut解析により求めるようにしてもよい。このNode-strut解析は、骨梁の連続性を二次元的に評価する方法であり、３個以上の骨梁の結合点をNode（Ｎｄ）（図19における白点）、他の骨梁と結合していない終末点をTerminus（Ｔｍ）（図19における黒点）と定義する。そして、図19に示すようにこれらをつなぐ骨梁の中心線（strut ）をＮｄＮｄ（ＮｄとＮｄとをつなぐstrut ）、ＮｄＴｍ（ＮｄとＴｍとをつなぐstrut ）、ＴｍＴｍ（ＴｍとＴｍとをつなぐstrut ）、ＣｔＮｄ（Ｃｔ（皮質骨）とＮｄとをつなぐstrut ）およびＣｔＴｍ（ＣｔとＴｍとをつなぐstrut ）に分類し、それぞれの長さを測定する。これらのstrut の長さとＮｄの数およびＴｍの数に基づいて以下のパラメータを定義する。
【０１０９】
（１）長さのパラメータ
すべてのstrut の長さの和（ＴＳＬ）に対する各strut の長さ（％）
ＮｄＮｄ／ＴＳＬ、ＮｄＴｍ／ＴＳＬ、ＴｍＴｍ／ＴＳＬ、ＣｔＮｄ／ＴＳＬ、ＣｔＴｍ／ＴＳＬ
（２）単位面積当たりのstrut の長さ（ｍｍ／ｍｍ² ）（なお、ＴＶは全骨組織面積または（全骨組織面積−皮質骨面積）を示す）
ＴＳＬ／ＴＶ、ＮｄＮｄ／ＴＶ、ＮｄＴｍ／ＴＶ、ＴｍＴｍ／ＴＶ、ＣｔＮｄ／ＴＶ、ＣｔＴｍ／ＴＶ
（３）数のパラメータ
Ｎ．Ｎｄ／ＴＶ＝単位面積当たりのＮｄの数（／ｍｍ² ）
Ｎ．Ｔｍ／ＴＶ＝単位面積当たりのＴｍの数（／ｍｍ² ）
Ｎ．Ｎｄ／Ｎ．Ｔｍ＝Ｎｄの数とＴｍの数との比
この定義ではＮｄ関連のパラメータが高いほど骨梁の連続性が高く、Ｔｍ関連のパラメータが高いほど骨梁の連続性が低いことを示す。このNode-strut解析は骨梁の連続性を直接かつ簡便に計測できる方法として有用性が高いものである。
【０１１０】
なお、図１に示した実施形態の骨部画像処理装置は、骨梁強調手段50が骨部画像信号Ｓsub に対して強調処理することによって、骨梁の画像情報を取得するものであるが、本発明の骨部画像処理装置はこの形態に限るものではなく、図８のフローチャートに示すように、高エネルギー画像Ａを表す高エネルギー画像信号Ｓａ、若しくは低エネルギー画像Ｂを表す低エネルギー画像信号Ｓｂに対して、または図９のフローチャートに示すように高エネルギー画像信号Ｓａと低エネルギー画像信号Ｓｂとを重み付け加算した加算信号に対して、骨梁強調手段50が強調処理することによって、骨梁の画像情報を取得する構成を採用することもできる。
【０１１１】
また、図８および図９に示すフローチャートにおいても、図15に示すフローチャートと同様に、骨梁の画像情報に基づいて骨梁構造の指標値を算出するようにしてもよい。さらに、算出された骨梁の指標値と骨密度の数値情報とに基づいて、被検体の骨部組織の状態を判定装置において判定するようにしてもよい。
【０１１２】
また骨密度解析手段30により求められた骨成分の密度に関する数値情報と骨梁強調手段50により求められた骨梁の形態情報とは、同一の出力媒体に出力してもよいし、各別の出力媒体に出力してもよい。
【図面の簡単な説明】
【図１】本発明の骨部画像処理装置の一実施形態を示すブロック図
【図２】図１に示した骨部画像処理装置の作用の流れを示すフローチャート
【図３】高エネルギー画像Ａ、低エネルギー画像Ｂおよびエネルギーサブトラクション処理手段により生成される骨部画像Ｃを示す図
【図４】骨塩レファレンスの一例を示す図
【図５】骨密度解析手段30により作成される濃度・骨塩量変換テーブルの一例を示す図
【図６】放射線画像における骨梁（骨粗鬆症の進行）を示す模式図
【図７】スケルトン処理後の骨梁（骨粗鬆症の進行）を示す模式図
【図８】本発明の骨部画像処理装置の他の実施形態の作用の流れを示すフローチャート（その１）
【図９】本発明の骨部画像処理装置の他の実施形態の作用の流れを示すフローチャート（その２）
【図１０】モーフォロジー処理の基本的な作用を示す図、（Ａ）ダイレーション（dilation）処理、（Ｂ）エロージョン（erosion ）処理、（Ｃ）オープニング（opening ）処理、（Ｄ）クロージング（closing ）処理
【図１１】モーフォロジー処理に用いる構造要素Ｂi の一例を示す図
【図１２】クロージング処理を具体的に説明する図
【図１３】種々の図形とその骨格（skeleton：スケルトン）を示す図
【図１４】スケルトン処理を具体的に説明する図
【図１５】本発明の骨部画像処理装置の他の実施形態の作用の流れを示すフローチャート（その３）
【図１６】 Star volume 法を説明するための図
【図１７】被検体の骨部組織の状態を判定する方法を説明するための図（その１）
【図１８】被検体の骨部組織の状態を判定する方法を説明するための図（その２）
【図１９】 Node-strut解析を説明するための図
【符号の説明】
10 エネルギーサブトラクション処理手段
30 骨密度解析手段
50 骨梁強調手段
100 骨部画像処理装置[0001]
BACKGROUND OF THE INVENTION
The present invention relates to a method and apparatus for acquiring information on bone components in a bone part such as a human body, and more specifically, a bone part for acquiring numerical information such as density of bone components useful for diagnosis of osteoporosis and the like and morphological information such as an image of trabecular bone The present invention relates to an image processing method and apparatus.
[0002]
[Prior art]
Bone mineral quantification, that is, quantitative measurement of the amount of calcium in bone, is useful for diagnosis of fracture prevention.
[0003]
In other words, the amount of bone mineral is determined by the density of the trabecular bone trabeculae that forms the inside of the bone, that is, the bone density. Therefore, if the bone density is sparse, the density of the bone shadow in the bone image increases. If the bone density is high, the density of the shadow of the bone part in the bone part image is low.
[0004]
Therefore, knowing the minute changes in calcium in bone enables early detection of osteoporosis and has an important effect in preventing fractures.
[0005]
Therefore, conventionally, MD method (Microdensitometry), SPA method (Single Photon Absorptiometry), DPA method (Dual Photon Absorptiometry), QDR method (Quantitative Digited Radiography), QCT method (Quantitative Computer Tomography), DQCT method (Dual energy Quantitative Computer Tomography) Numerous bone mineral quantification methods have been proposed and implemented.
[0006]
However, these techniques have respective drawbacks such as inability to perform bone mineral quantification of vertebrae that directly represent osteoporosis, a large-scale apparatus, and an increased radiation exposure.
[0007]
Therefore, the applicant of the present application has proposed a bone mineral quantitative analysis method using energy subtraction processing in Japanese Patent Application Laid-Open No. 4-11473.
[0008]
The method using the energy subtraction process is to irradiate a subject including soft tissue and bone tissue and transmit the transmitted radiation to two or more storage phosphor sheets or recording sheets such as X-ray films. In addition, the bone part is recorded by subtracting the digital image signal obtained by reading from each recording sheet between the corresponding pixels of each image so as to erase the soft tissue. In energy subtraction processing for obtaining a difference signal that forms an image of only tissue (bone part image), when recording a radiographic image of a subject, a bone mineral imitating a human bone in which a known bone mineral amount has been changed stepwise in advance. By recording the reference with the subject and comparing the bone tissue reference density with the bone mineral reference density on the bone image A method of quantifying the amount of salt.
[0009]
[Problems to be solved by the invention]
However, although the method disclosed in JP-A-4-11473 can accurately measure the density of bone components, it is required to provide information useful for diagnosis in osteoporosis diagnosis. .
[0010]
That is, even if only numerical information called bone density is presented, it is difficult to grasp the state of the trabecular bone by itself.
[0011]
The present invention has been made in view of the above circumstances, and provides a bone image processing method and apparatus capable of easily recognizing the state of trabecular bone in bone tissue and acquiring information useful for diagnosis. It is the purpose.
[0012]
[Means for Solving the Problems]
The bone image processing method and apparatus according to the present invention obtains numerical information related to the density of bone components by performing energy subtraction processing on a radiographic image, and performs trabecular enhancement processing on the radiographic image. It is characterized by acquiring trabecular morphology information, and by acquiring numerical information and morphological information, it is easier to recognize the state of trabecular bone, and is particularly useful for accurately diagnosing osteoporosis.
[0013]
That is, the first bone image processing method of the present invention is based on two radiographic images having different radiation energy distributions for energy subtraction processing with a subject including a soft tissue and a bone tissue as subjects. A bone image processing method for obtaining information on the density of bone components in the bone tissue,
By performing energy subtraction processing on the two radiation images, a bone image in which an image of bone tissue in the subject is extracted or emphasized is generated,
Based on the bone image, to obtain numerical information about the density of the bone component,
By performing a trabecular emphasis process on the bone part image, the trabecular shape information is acquired.
[0014]
In this case, based on the shape information of the trabecular bone, an index value representing trabecular information may be calculated, and further, based on the index value representing the trabecular information and numerical information on the density of the bone component, The state of the bone tissue of the subject may be determined.
[0015]
Here, as the trabecular enhancement process, it is preferable to apply a skeleton process based on a morphological operation.
[0016]
Here, morphological operations (hereinafter also referred to as morphological processing) will be described.
[0017]
Morphology processing is processing based on a morphology (Morphology) algorithm that selectively extracts only a specific image portion such as an abnormal shadow from an original image, and is particularly characteristic in breast cancer. However, the target image is not limited to the microcalcification image in such a mammogram.
[0018]
This morphological process is a process using the structural element B corresponding to the size and shape of the image portion to be extracted, is not easily affected by complex background information, and the extracted image is not distorted. There are features such as.
[0019]
In other words, this method can detect the calcified image such as the size, shape, and density distribution of the calcified image while maintaining it better than general differentiation processing.
[0020]
Hereinafter, an example in which the outline of the morphological process is applied to detection of a microcalcification image in a mammogram will be described.
[0021]
(Basic morphological operations)
The morphological processing is generally developed as a set theory in an N-dimensional space, but will be described for a two-dimensional gray image for intuitive understanding.
[0022]
The grayscale image is regarded as a space in which a point at coordinates (x, y) has a height corresponding to the density value f (x, y). Here, the density value f (x, y) is a signal with a high luminance and high signal level that becomes a larger image signal value as the density is lower (the luminance is higher when displayed on the CRT).
[0023]
First, for the sake of simplicity, a one-dimensional function f (x) corresponding to the cross section of the two-dimensional gray image is considered. The structural element g used for the morphological operation is a symmetric function symmetric about the origin as shown in the following equation (3).
[0024]
[Equation 3]

[0025]
It is assumed that the value is 0 in the domain and the domain G is the following formula (4).
[0026]
[Expression 4]

[0027]
At this time, the basic form of the morphological operation is a very simple operation as shown in the equations (5) to (8).
[0028]
[Equation 5]

[0029]
That is, the dilation processing is performed at the maximum within a range of ± m (a value determined according to the structural element B and corresponding to the mask size in FIG. 10) with the pixel of interest at the center. On the other hand, the erosion process is a process for searching for a minimum value within a range of ± m centered on the pixel of interest (see FIG. 4A). (Refer figure (B)). The opening process is a process for performing dilation after the erosion process, that is, a process for searching for the maximum value after searching for the minimum value, and the closing process is a process for performing erosion process after the dilation process. That is, this corresponds to a process of searching for the minimum value after searching for the maximum value.
[0030]
That is, the opening process smoothes the density curve f (x) from the low luminance side, and removes the convex density fluctuation portion (a portion having a higher luminance than the surrounding portion) that fluctuates in a spatially narrower range than the mask size 2 m. (Refer to FIG. 2C).
[0031]
On the other hand, the closing process smoothes the density curve f (x) from the high luminance side, and removes the concave density fluctuation portion (the portion whose luminance is lower than the surrounding portion) that fluctuates in a spatially narrower range than the mask size 2 m. (See FIG. 4D).
[0032]
In the case where the structural element g is not symmetric with respect to the origin, the dilation operation shown in Equation (5) is called Minkowski sum, and the erosion operation shown in Equation (6) is called Minkowski difference.
[0033]
Here, in the case of a signal having a high density and high signal level, which has a larger value as the density is higher, the magnitude relationship is reversed as compared with the case where the image signal value of the density value f (x) is a high luminance and high signal level. Therefore, the dilation processing for the high density high signal level signal and the erosion processing for the high brightness high signal level ((B) in the figure) are the same, and the erosion processing for the high density high signal level signal and the high brightness high signal level. Is the same as the dilation processing (FIG. (A)), and the opening processing for the signal of high density and high signal level is the same as the closing processing (FIG. (D)) of high luminance and high signal level. The closing process for the signal level signal and the opening process for the high luminance and high signal level ((C) in the figure) are the same.
[0034]
In this section, a case of an image signal (brightness value) having a high brightness and a high signal level will be described.
[0035]
(Application to calcified shadow detection)
For the detection of the calcified shadow, a difference method in which a smoothed image is removed from the original image can be considered. Since it is difficult to distinguish between calcified shadows and elongated non-calcified shadows (such as mammary glands, blood vessels, and mammary support tissues) with a simple smoothing method, Obata et al. At Tokyo University of Agriculture and Technology used multiple structural elements. A morphological filter represented by the following formula (9) based on the opening process has been proposed ("Extraction of microcalcifications by morphological filter using multiple structural elements" IEICE Transactions D-II Vol. .J75-D-II No.7 P1170-1176 July 1992, "Basics of Morphology and Its Application to Mammogram Processing" MEDICAL IMAGING TECHNOLOGY Vol.12 No.1 January 1994, etc.).
[0036]
[Formula 6]

[0037]
Here, Bi (i = 1, 2,..., N) has a linear size of m pixels and n (for example, in the case shown in FIG. 11, nine pixels are in four directions and m = 9, n = 4). These structural elements (hereinafter collectively referred to as multiple structural elements in the direction of m pixels and n). If the structuring element Bi is set to be larger than the calcified shadow to be detected, the above-mentioned processing by the opening calculation will produce a signal change portion (image portion where the signal fluctuates in a spatially narrow range) finer than the structuring element Bi. Thus, the calcified shadow, which is a convex portion having a luminance value larger than that of the surrounding area, is removed. On the other hand, if the length of the non-calcified shadow such as the shadow of the elongated mammary gland is longer than that of the structural element Bi and the inclination coincides with one of the four structural elements Bi, the opening process (the expression (9)) It remains as it is even after the operation of the second term). Therefore, by removing the smoothed image (image from which only the calcified shadow has been removed) obtained by the opening process from the original image f, an image including only a small calcified shadow can be obtained. This is the idea of equation (9).
[0038]
As described above, in the case of a signal having a high density and a high signal level, the calcified shadow has a lower density value than the surrounding image portion, and the calcified shadow has a concave shape with a smaller density value than the surrounding portion. Since this is a signal change portion, the closing process is applied instead of the opening process, and the expression (10) is applied instead of the expression (9).
[0039]
[Expression 7]

[0040]
The closing process of Expression (10), which is an example of the morphological operation, will be specifically described.
[0041]
That is, according to the morphological calculation for the density value Sorg, which is an image signal having a high density and high signal level, for example, image data having a distribution of density values Sorg as shown by the solid line in FIG. By performing maximum value processing (dilation processing) on the linear three-pixel structure element B as shown in (2), the density value S of a certain target pixel is obtained._iIs the maximum value S among the three pixels (determined by the structural element B) that are adjacent to each other around the target pixel._{i + 1}S adopted_iConverted to ′. By performing this calculation for all the pixels, the signal is converted into a maximum value signal having a distribution of density values Sorg ′ and indicated by a broken line in FIG.
[0042]
Next, considering the minimum value processing (erosion processing) by the structural element B with respect to the maximum value signal obtained by this maximum value processing, the maximum value signal of the target pixel indicated by the broken line in FIG. S_i′ Is the minimum value S among the three pixels adjacent to each other with the target pixel as the center._i-1S adopting '_i″ (= S_i). By performing this calculation for all the pixels, the distribution of the minimum value signal Sorg ″ after the maximum value processing is indicated by a one-dot chain line in FIG. 1 (1). The image signal indicated by the one-dot chain line is In contrast to the original solid line image data, the image portion where the signal fluctuates in a spatial range narrower than that of the structural element B disappears, and the signal value changing portion which fluctuates in a spatial range wider than that of the structural element B This means that the image portion and the image portion having no fluctuation remain in their original shapes, that is, the above processing (closing processing) is a processing for smoothing the image density distribution from the high density side. Acts as.
[0043]
The value Smor obtained by subtracting from the original image signal Sorg the value obtained by the closing process (the value obtained by further performing the minimum value process after performing the maximum value process on Sorg) is the above-mentioned closing process. Represents an image portion which is a changed portion of the signal value which fluctuates in a spatially narrow range erased by.
[0044]
Here, the image signal originally has a position (x, y) which is a two-dimensional element and a signal value f (x, y) which is a third dimension element. However, in the above description, it is easy to understand. Therefore, the one-dimensional image signal distribution curve that appears in a predetermined section of the image developed two-dimensionally has been described.
[0045]
Therefore, in actuality, it is necessary to apply the above description to the two-dimensional image, and the reason why the multiple structural elements are used is to correspond to the two-dimensional image.
[0046]
Next, skeleton processing based on morphological operations will be described.
[0047]
The skeleton processing is generally processing for extracting a skeleton of a figure, and the skeleton can be regarded as a set of centers of disks inscribed in the figure. That is, for example, the skeleton of each figure (shown by a solid solid line) shown in FIGS. 13A to 13E is indicated by a thick solid line.
[0048]
Hereinafter, the case where this skeleton process is performed by the morphological operation described above will be described. In this case, the skeleton process can be expressed by the following formula (1) or (2).
[0049]
[Expression 1]

[0050]
[Expression 2]

[0051]
Here, the expression (1) and the expression (2) are due to the difference between whether the image is expressed as an image signal having a high density and a high signal level or as an image signal having a high luminance and a high signal level. When extracting the skeleton of the low density (high luminance) image portion from the image represented by the image signal of the high density signal level, the expression (1) is applied, while on the other hand, the skeleton of the high density high signal level image signal is applied. When extracting the skeleton of the low luminance (high density) image portion from the obtained image, Equation (2) is applied, and there is no substantial difference in the operation itself.
[0052]
For example, on a negative film (high density and high signal level), the density of the bone part is lower than that of other image parts, the density of the part where the trabecular bone is present, and the density of the part where the bone is not present is high. Therefore, since it corresponds to performing a skeleton process on a trabecular bone having a lower density than the surrounding area, the equation (1) may be applied.
[0053]
Here, the structural element B in the formula (1) is a circle having a radius r, and a state in which the skeleton processing is applied to the figure shown in FIG. In the figure shown in FIG. 14, the outer region is a portion having a higher density than the contour thereof, and the inner region is a portion having a lower density.
[0054]
First, the figure is subjected to erosion processing by the structural element B in the upper part of the figure. When λ = 0 (zero erosion processing by the structural element B), the figure does not change at all.
[0055]
When λ = 1 (one erosion process by the structural element B), the figure is embedded inside by the radius r of the structural element B.
[0056]
When λ = 2 (twice erosion processing by the structural element B), the portion protruding from the circle of the figure disappears completely.
[0057]
By repeating the same operation, the figure becomes only a circle having a radius r or less with λ = N−1 (N−1 erosion processing by the structural element B).
[0058]
On the other hand, in the middle part of the figure, an opening process by the structural element B is further performed on the image subjected to the erosion process of each time (λ = 0, 1, 2,..., N−1, N) by the structural element B. It is a figure.
[0059]
A figure obtained by subtracting the figure in the middle part from the figure in the upper part in the figure in accordance with the number of times of processing is the lower part in the figure.
[0060]
It can be seen that when λ = 1, the skeleton element of the portion protruding from the circle of the original graphic is extracted, and when λ = N−1, the skeleton element of the circle of the original graphic is extracted.
[0061]
The expression (1) means that the erosion process is performed on the original figure, the opening process is further performed, the number of processes is subtracted, and the union of the results is obtained.
[0062]
In Expression (2), the level of density is effective for extracting the skeleton element from the figure reversed from Expression (1). Dilation processing is performed on the original graphic, and then closing processing is performed. This means that the skeleton elements are extracted by subtracting them in correspondence with each other and obtaining the union of the results.
[0063]
As the union in the above formulas (1) and (2), only the union having a relatively large value of n1, that is, λ = 2, 3, 4, 5 except for the case of n1 = 0, 1, for example. By adopting a union set such as, and displaying the skeletal elements obtained in this case, changes in trabeculae can be more easily seen, which is preferable. This is because noise components having a very high spatial frequency are also extracted, particularly when λ is 0 or 1, and by removing these noise components from the union set, an image with higher interpretation performance can be obtained. Because it can. The same applies to the following inventions.
[0064]
The second bone image processing method of the present invention is based on two radiographic images having a subject including a soft tissue and a bone tissue and having different radiation energy distributions for energy subtraction processing. A bone image processing method for obtaining information on the density of bone components in the tissue,
By performing energy subtraction processing on the two radiation images, a bone image in which an image of bone tissue in the subject is extracted or emphasized is generated, and based on the bone image, the bone component Get numerical information about density,
The trabecular shape information is obtained by performing a trabecular enhancement process on at least one of the two radiographic images or an addition image obtained by weighting and adding the two radiographic images in correspondence with pixels. It is characterized by acquiring.
[0065]
That is, the first bone image processing method of the present invention performs the trabecular enhancement process based on the bone image after energy subtraction processing of two radiographic images for energy subtraction processing. On the other hand, in the second bone image processing method of the present invention, prior to performing the energy subtraction process, at least one of the two radiographic images for the energy subtraction process, or the two radiographic images are weighted with corresponding pixels. A trabecular enhancement process is performed on the added image obtained by the addition.
[0066]
Also in the second bone image processing method of the present invention, an index value representing a trabecular structure may be calculated based on the trabecular shape information, as in the first bone image processing method. Further, the condition of the bone tissue of the subject may be determined based on the index value representing the trabecular structure and the numerical information related to the density of the bone component.
[0067]
In each of the above inventions, as a method for obtaining numerical information related to the density of bone components, in addition to the subject, a bone mineral reference whose bone mineral content is known, that is, a reference imitating a human bone, A radiographic image is generated by adding to the subject an object created by changing in stages (the amount of bone mineral in each stage is known), and the density of the bone mineral reference image appearing in the bone part image and the known bone A method of obtaining a concentration / bone salt amount conversion table from the salt amount and acquiring the bone portion concentration in the bone image of the subject based on the concentration / bone salt amount conversion table (e.g. For example, see Kaihei 4-11473).
[0068]
The bone part image processing apparatus of the present invention is an apparatus for carrying out the bone part image processing method of the present invention, and is a radiation energy for energy subtraction processing for a subject including a soft tissue and a bone part tissue. A bone image processing apparatus that acquires information on the density of bone components in the bone tissue based on two or more radiation images having different distributions,
Energy subtraction processing means for performing energy subtraction processing for extracting or enhancing a desired bone portion of an image represented by the image signal based on two predetermined image signals;
Based on the image signal of the bone part, bone density analysis means for obtaining numerical information on the density of the bone component;
Each image signal representing the two radiation images before the energy subtraction process, the image signal of the bone part after the energy subtraction process, or the addition obtained by weighted addition of the image signals representing the two radiation images A trabecular enhancement means for emphasizing the trabecular bone with respect to the image signal is provided.
[0069]
Further, the bone image processing apparatus of the present invention includes index value calculation means for calculating an index value representing the trabecular structure based on the shape information of the trabecular bone emphasized by the trabecular emphasis means. In addition, a determination unit that determines the state of the bone tissue of the subject based on the index value representing the trabecular structure and the numerical information related to the density of the bone component may be provided.
[0070]
Here, as the trabecular bone emphasizing means, morphological operation means for applying a skeleton process based on the morphological operation represented by the formula (1) or (2) can be applied.
[0071]
When the image signal to which the morphological operation is applied is represented by an image signal having a high density and a high signal level, Expression (1) is applied, and the image signal to which the morphological operation is applied has a high luminance and high signal level. If it is represented by an image signal, Equation (2) is applied.
[0072]
In an image represented by an image signal having a high density and a high signal level, the trabecular bone is represented as a low density image portion. In an image represented by an image signal having a high intensity and a high signal level, the trabecular bone has a high luminance. This is because it is expressed as an image part.
[0073]
As the union in the above formulas (1) and (2), only the union having a relatively large value of n1, that is, λ = 2, 3, 4, 5 except for the case of n1 = 0, 1, for example. By adopting a union set such as, and displaying the skeletal elements obtained in this case, changes in trabeculae can be more easily seen, which is preferable.
[0074]
Further, as the bone density analysis means, the bone mineral reference recorded in the radiographic image together with the subject, the bone mineral reference of which the bone mineral amount is known, the density of the bone mineral reference image appearing in the bone part image, and the known Based on a density / bone mineral content conversion table created from the bone mineral content, the bone density in the bone image of the subject is converted into the bone mineral content to obtain numerical information on the density of the bone component Can be applied.
[0075]
【The invention's effect】
The bone image processing method and apparatus according to the present invention obtains a bone image by performing energy subtraction processing on a radiographic image, acquires numerical information relating to the density of bone components based on the bone image, and also provides a radiographic image. Or, it is easy to visually recognize the state of trabecular bone by acquiring trabecular shape information by performing trabecular emphasis processing on the bone part image, and is particularly useful for diagnosis of osteoporosis Can be provided.
[0076]
In particular, by calculating an index value that represents the trabecular structure based on the trabecular morphology information, the trabecular state can be obtained as a numerical value, thereby quantitatively obtaining information useful for the diagnosis of osteoporosis. Can do.
[0077]
DETAILED DESCRIPTION OF THE INVENTION
Hereinafter, specific embodiments of the bone image processing apparatus of the present invention will be described with reference to the drawings.
[0078]
FIG. 1 is a block diagram showing an embodiment of a bone image processing apparatus of the present invention, and FIG. 2 is a diagram showing a flow of operation of the bone image processing apparatus shown in FIG.
[0079]
The illustrated bone image processing apparatus 100 is a high-energy image A for energy subtraction processing (see FIG. 3) using a subject including soft tissues (A2, B2) and bone tissues (A1, B1) as subjects. And a low energy image B, an energy subtraction processing means 10 for generating a bone image C in which the bone tissue C1 is emphasized, and a bone image signal (energy subtraction image signal) Ssub representing the bone image C The bone density analysis means 30 for obtaining numerical information on the density of the bone component in the bone tissue C1 and the trabecular enhancement means 50 for emphasizing the trabecular bone with respect to the bone image signal Ssub. is there.
[0080]
Here, the high energy image A and the low energy image B are images having different energy distributions obtained by one photographing by so-called one-shot energy subtraction (see Japanese Patent Laid-Open No. 59-83486, etc.). In addition to the image of the subject, images A3 and B3 of the bone mineral reference 31 (see FIG. 4) taken together with the subject are recorded.
[0081]
Bone mineral reference 31 is the amount of bone mineral, that is, CaCO_ThreeIs a structure in which the content (wt%) is changed in stages and arranged in different sections d1, d2, d3, d4, d5, d6, and CaCO in all sections_ThreeThe content of each is known.
[0082]
Further, the energy subtraction processing means 10 generates the soft tissue B2 and the high level of the low energy image B based on the input high energy image signal Sa representing the high energy image A and low energy image signal Sb representing the low energy image B. A weight difference between the two images (difference signal = energy subtraction image signal) Ssub is calculated after weighting both images so that the density of the energy image A matches the soft tissue A2. As a result, the energy subtraction image signal Ssub represents a bone image in which the bone tissue C1 is emphasized.
[0083]
The bone density analyzing means 30 obtains the density of each section d1, d2, d3, d4, d5, d6 of the image C3 of the bone mineral reference 31 appearing in the bone image C, and the sections d1, d2 , D3, d4, d5, d6 and the corresponding bone mineral density for each section d1, d2, d3, d4, d5, d6, and corresponding concentrations as shown in the graph of FIG.・ Create bone mineral content conversion table. Then, the density of the bone tissue C1 appearing in the bone image C is obtained, and the obtained density is converted into the bone mineral content according to the created density / bone mineral content conversion table, whereby numerical information regarding the density of the bone component is obtained. Is what you want.
[0084]
The trabecular bone emphasizing means 50 emphasizes the trabecular bone of the bone tissue C1 in the bone image C by performing a skeleton process based on a morphological operation represented by the following formula (1).
[0085]
[Expression 1]

[0086]
Next, the operation of the bone image processing apparatus 100 of this embodiment will be described.
[0087]
The energy subtraction processing means 10 performs an energy subtraction process based on the following formula (11) from the input high energy image signal Sa representing the high energy image A and low energy image signal Sb representing the low energy image B, and A partial image signal Ssub is obtained.
[0088]
Ssub = K log Sa−L log Sb + M (11)
(K, L, and M represent coefficients)
The obtained bone image signal Ssub represents the bone image C in which the soft tissues A2 and B2 are erased and the bone tissues A1 and B1 are relatively emphasized.
[0089]
Next, the bone part image signal Ssub representing the bone part image C obtained in this way is inputted to the bone density analyzing means 30, and the bone density analyzing means 30 is used for the bone mineral reference 31 appearing in the bone part image C. The density (image signal) of each section d1, d2, d3, d4, d5, d6 of the image C3 is obtained.
[0090]
And the concentration for each section d1, d2, d3, d4, d5, d6 and the known bone mineral content (CaCO for each section d1, d2, d3, d4, d5, d6)_ThreeThe concentration / bone mineral amount conversion table (see FIG. 5) is created in correspondence with the amount. Thereafter, the bone portion image signal Ssub is converted into the bone mineral amount according to the density / bone mineral amount conversion table to obtain numerical information relating to the density of the bone component in the bone tissue C1.
[0091]
The obtained numerical information is output to an output medium such as a display device and used for diagnosis.
[0092]
On the other hand, the bone image signal Ssub representing the bone image C is also input to the trabecular bone emphasizing means 50.
[0093]
The trabecular bone emphasizing means 50 performs skeleton processing based on the morphological operation represented by the above formula (1) for the inputted bone image signal Ssub, and emphasizes the trabeculae of the bone tissue C1 in the bone image C. To process. The structural element used for the morphological calculation is set to a size suitable for extracting the trabecular skeleton (specifically, a size corresponding to the thickness of the trabecular bone).
[0094]
The trabecular bone is spongy that extends vertically and horizontally inside the bone. In the radiographic image, as shown in FIG. 6, the spongy material extending vertically and horizontally is displayed in multiple layers. As shown in FIG. 7, an image in which only the spongy skeleton is overlapped is formed.
[0095]
An image in which only the bone trabecula skeleton is extracted is output to an external display means or the like.
[0096]
Thus, according to the bone image processing apparatus of the present embodiment, the density of the bone component can be quantitatively obtained as numerical information, and the state of the trabecular bone can also be obtained as visual image information. Therefore, it is very effective in diagnosing the condition of trabecular bone and can provide information useful for diagnosis of osteoporosis.
[0097]
The bone part image processing apparatus of the embodiment shown in FIG. 1 acquires trabecular image information in the trabecular emphasis means 50. As shown in the flowchart of FIG. Based on the image information of the trabecular bone obtained in 50, an index value indicating the trabecular structure is calculated, and the index value and numerical information of the bone density are input to the determination device, and the bone tissue of the subject is calculated. The state may be determined. Hereinafter, calculation of the index value will be described.
[0098]
A specific example of the method for calculating the index value of the trabecular structure is the Star volume method. As shown in FIG. 16, this Star volume method is an average of the marrow space star volume (Vm) that means the average of the marrow cavity volume in a range that can be seen without being blocked by the trabecular bone from all points in the marrow cavity. Similarly, two types of trabecular star volumes (Vt) are defined which mean the average of the volume in the range from the point in the trabecular bone to the end of the trabecular bone in all directions. In FIG. 16, the shaded portion indicates the trabecular bone and the other portion indicates the bone marrow cavity. Star volume is the size of the trabecular bone in the bone marrow cavity by devising the sampling method of the specimen.^ThreeOr μm^ThreeIt is considered as a three-dimensional index without a gap expressed as a three-dimensional value. Vm is small when trabecular continuity is high, and is large when trabecular disappearance and drilling are many. Conversely, Vt is large when trabecular continuity is high, and when trabecular disappearance and drilling are many. Becomes smaller.
[0099]
Vmi at any point i in the bone marrow cavity is defined by the following equation (12).
[0100]
Vmi = π / 3 × l_0m ^Three                  (12)
However, l₀Indicates the length of continuous bone marrow cavity in any direction centered on point i. Also, l_0m ^ThreeIs l in all directions radially about the point i₀ ^ThreeIt represents the average value when.
[0101]
Next, Vtj at an arbitrary point j on the trabecular bone is defined by the following equation (13).
[0102]
Vtj = π / 3 × Σl₁ ^Four/ Σl₁         (13)
However, l₁Indicates the length of continuous trabecular bone in any direction centered on point j. Also, l₁Is obtained in all directions radially around the point j, and Σ represents obtaining the sum in all directions. In this embodiment, Vmi and Vtj are calculated for each sampling point, and the average values are Vm and Vt.
[0103]
Note that either Vm or Vt may be used as the index value of the trabecular structure, but Vt is used in the present embodiment.
[0104]
After obtaining trabecular image information, the value of Vt is obtained based on the above equation (13), and the value of Vt is input to the determination apparatus together with the numerical information of the bone density.
[0105]
In the determination apparatus, for example, as shown in FIG. 17, three-dimensional coordinates with the subject's age, bone density value, and Vt as axes are formed, and the input result is at any position within the three-dimensional coordinates. The state of the bone tissue is determined according to whether it exists. That is, if the age, bone density value, and Vt of the subject input to the determination device are at the position indicated by the point P1, it is determined that the state is A, and if it is at the position indicated by the point P2, it is determined that the state is B. If it is at the position indicated by the point P3, it is determined that it is in the state C.
[0106]
Thus, by obtaining the trabecular index value based on the image information of the trabecular bone, it is possible to quantitatively obtain the state of the trabecular bone as well as the density of the bone component. It is very effective for diagnosis, can provide information useful for diagnosis of osteoporosis in particular, and can automatically determine the condition of the bone tissue of the subject.
[0107]
In the embodiment shown in FIG. 15, as shown in FIG. 17, three-dimensional coordinates with the subject's age, bone density value, and Vt as axes are formed to determine the state of the bone tissue of the subject. However, as shown in FIG. 18A, provisional determinations 1 to 3 are obtained based on the age and bone density value of the subject, and as shown in FIG. The state of the bone tissue of the subject may be determined with reference to a table representing the relationship with the value of Vt.
[0108]
In the embodiment shown in FIG. 15, Vt is obtained as an index value, but Vm may be obtained as an index value, and further, the index value may be obtained by node-strut analysis. This node-strut analysis is a method for two-dimensional evaluation of trabecular continuity. The connection point of three or more trabeculae is defined as Node (Nd) (white point in FIG. 19) and other trabeculae. An unconnected end point is defined as Terminus (Tm) (black point in FIG. 19). Then, as shown in FIG. 19, the center line (strut) of the trabecular bone connecting these is NdNd (strut connecting Nd and Nd), NdTm (strut connecting Nd and Tm), TmTm (connecting Tm and Tm). strt), CtNd (strut connecting Ct (cortical bone) and Nd) and CtTm (strut connecting Ct and Tm), and measuring the lengths thereof. The following parameters are defined based on the length of strut, the number of Nd, and the number of Tm.
[0109]
(1) Length parameter
Length of each strut (%) relative to the sum of all strut lengths (TSL)
NdNd / TSL, NdTm / TSL, TmTm / TSL, CtNd / TSL, CtTm / TSL
(2) Strut length per unit area (mm / mm²(TV indicates total bone tissue area or (total bone tissue area-cortical bone area))
TSL / TV, NdNd / TV, NdTm / TV, TmTm / TV, CtNd / TV, CtTm / TV
(3) Number parameters
N. Nd / TV = number of Nd per unit area (/ mm²)
N. Tm / TV = number of Tm per unit area (/ mm²)
N. Nd / N. Tm = ratio of the number of Nd to the number of Tm
This definition indicates that the higher the Nd related parameter, the higher the trabecular continuity, and the higher the Tm related parameter, the lower the trabecular continuity. This node-strut analysis is highly useful as a method that can directly and simply measure the continuity of trabecular bone.
[0110]
In the bone image processing apparatus according to the embodiment shown in FIG. 1, the trabecular emphasis unit 50 obtains image information of the trabecular bone by performing an emphasis process on the bone image signal Ssub. The bone image processing apparatus of the present invention is not limited to this form, and as shown in the flowchart of FIG. 8, a high energy image signal Sa representing a high energy image A or a low energy image signal Sb representing a low energy image B. As shown in the flowchart of FIG. 9, the trabecular emphasis means 50 performs an emphasis process on the addition signal obtained by weighted addition of the high energy image signal Sa and the low energy image signal Sb. A configuration for acquiring image information can also be adopted.
[0111]
8 and 9, the trabecular structure index value may be calculated based on the trabecular image information as in the flowchart shown in FIG. Furthermore, the state of the bone tissue of the subject may be determined by the determination device based on the calculated trabecular index value and the numerical information of the bone density.
[0112]
Further, the numerical information regarding the density of the bone component obtained by the bone density analyzing means 30 and the trabecular shape information obtained by the trabecular emphasis means 50 may be output to the same output medium, You may output to an output medium.
[Brief description of the drawings]
FIG. 1 is a block diagram showing an embodiment of a bone image processing apparatus of the present invention.
FIG. 2 is a flowchart showing a flow of operation of the bone image processing apparatus shown in FIG.
FIG. 3 is a view showing a bone image C generated by a high energy image A, a low energy image B, and energy subtraction processing means.
FIG. 4 is a diagram showing an example of a bone mineral reference
FIG. 5 is a diagram showing an example of a concentration / bone mineral amount conversion table created by the bone density analyzing means 30.
FIG. 6 is a schematic diagram showing trabecular bone (progress of osteoporosis) in a radiographic image.
FIG. 7 is a schematic diagram showing trabecular bone (progress of osteoporosis) after skeleton treatment.
FIG. 8 is a flowchart (part 1) showing the flow of operation of another embodiment of the bone image processing apparatus of the present invention.
FIG. 9 is a flowchart (part 2) showing the flow of operation of another embodiment of the bone image processing apparatus of the present invention.
FIG. 10 is a diagram showing the basic operation of morphological processing, (A) dilation processing, (B) erosion processing, (C) opening processing, and (D) closing. processing
FIG. 11 is a diagram showing an example of a structural element Bi used for morphological processing.
FIG. 12 is a diagram for specifically explaining the closing process;
FIG. 13 is a diagram showing various figures and their skeletons.
FIG. 14 is a diagram for specifically explaining skeleton processing;
FIG. 15 is a flowchart (part 3) showing the flow of operation of another embodiment of the bone image processing apparatus of the present invention;
FIG. 16 is a diagram for explaining the Star volume method.
FIG. 17 is a diagram for explaining a method for determining the state of a bone tissue of a subject (part 1);
FIG. 18 is a diagram for explaining a method for determining the state of a bone tissue of a subject (part 2);
FIG. 19 is a diagram for explaining node-strut analysis.
[Explanation of symbols]
10 Energy subtraction processing means
30 Bone density analysis means
50 Trabecular enhancement
100 bone image processing device

Claims (12)

Based on two radiographic images represented by high-density and high-signal level image signals in which the radiation energy distribution for energy subtraction processing is different from each other, and the subject includes a soft tissue and a bone tissue. A bone image processing method for obtaining information on the density of bone components in the tissue,
By performing energy subtraction processing on the two radiation images, a bone image in which an image of bone tissue in the subject is extracted or emphasized is generated,
Based on the bone image, obtain numerical information on the density of the bone component,
By performing a trabecular enhancement process using a skeleton process based on a morphological calculation represented by the following formula (1) for the bone part image, the morphology information of the trabecular bone is obtained,
Based on the morphology information of the trabecular bone, an index value representing the trabecular structure is calculated,
A bone image processing method, comprising: determining a state of a bone tissue of the subject based on numerical information related to the density of the bone component and an index value representing the trabecular structure.

Based on two radiographic images represented by high-density and high-signal level image signals in which the radiation energy distribution for energy subtraction processing is different from each other, and the subject includes a soft tissue and a bone tissue. A bone image processing method for obtaining information on the density of bone components in the tissue,
By performing energy subtraction processing on the two radiation images, a bone image in which an image of bone tissue in the subject is extracted or emphasized is generated,
Based on the bone image, obtain numerical information on the density of the bone component,
Based on a morphological operation represented by the following formula (1) with respect to at least one of the two radiographic images or an addition image obtained by weighting and adding the two radiographic images in correspondence with pixels. By obtaining the trabecular emphasis processing by using the skeleton processing to obtain the shape information of the trabecular bone,
Based on the morphology information of the trabecular bone, an index value representing the trabecular structure is calculated,
A bone image processing method, comprising: determining a state of a bone tissue of the subject based on numerical information related to the density of the bone component and an index value representing the trabecular structure.

3. The bone image processing method according to claim 1, wherein the value of n1 in the formula (1) is 2 or more. Based on two radiographic images represented by image signals of high luminance and high signal levels, each of which has a subject including a soft tissue and a bone tissue, the radiation energy distribution for energy subtraction processing being different from each other, and the bone A bone image processing method for obtaining information on the density of bone components in the tissue,
  By performing energy subtraction processing on the two radiation images, a bone image in which an image of bone tissue in the subject is extracted or emphasized is generated,
  Based on the bone image, obtain numerical information on the density of the bone component,
  By performing a trabecular enhancement process using a skeleton process based on a morphological operation represented by the following formula (2) for the bone part image, the morphology information of the trabecular bone is obtained,
  Based on the morphology information of the trabecular bone, an index value representing the trabecular structure is calculated,
  A bone image processing method, comprising: determining a state of a bone tissue of the subject based on numerical information relating to the density of the bone component and an index value representing the trabecular structure.

Based on two radiographic images represented by image signals of high luminance and high signal levels, each of which has a subject including a soft tissue and a bone tissue, the radiation energy distribution for energy subtraction processing being different from each other, and the bone A bone image processing method for obtaining information on the density of bone components in the tissue,
  By performing energy subtraction processing on the two radiation images, a bone image in which an image of bone tissue in the subject is extracted or emphasized is generated,
  Based on the bone image, obtain numerical information on the density of the bone component,
  Based on a morphological operation represented by the following formula (2) with respect to at least one of the two radiographic images, or an addition image obtained by weighting and adding the two radiographic images in correspondence with pixels. By applying trabecular enhancement using skeleton processing Obtaining the morphology information of the trabecular bone,
  Based on the morphology information of the trabecular bone, an index value representing the trabecular structure is calculated,
  A bone image processing method, comprising: determining a state of a bone tissue of the subject based on numerical information relating to the density of the bone component and an index value representing the trabecular structure.

6. The bone image processing method according to claim 4, wherein the value of n1 in the formula (2) is 2 or more. In addition to the subject, a bone mineral reference having a known bone mineral content is added to the subject to generate the radiation image,
A density / bone mineral amount conversion table is created from the density of the bone mineral reference image appearing in the bone part image and the known bone mineral quantity,
7. The numerical information is obtained by converting a bone density in the bone image of the subject into a bone mineral quantity based on the density / bone mineral quantity conversion table. The bone image processing method according to claim 1. Based on two radiographic images represented by high-density and high-signal level image signals in which the radiation energy distribution for energy subtraction processing is different from each other, and the subject includes a soft tissue and a bone tissue. A bone image processing apparatus for obtaining information on the density of bone components in the tissue,
Energy subtraction processing means for performing energy subtraction processing for extracting or enhancing a desired bone portion of an image represented by the image signal based on two predetermined image signals;
Based on the image signal of the bone part, bone density analysis means for obtaining numerical information on the density of the bone component;
An image signal representing at least one of the two radiographic images before the energy subtraction processing, an image signal representing an added image obtained by weighting and adding the two radiographic images in correspondence with pixels, or the energy A trabecular enhancement means for emphasizing a trabecular bone using a skeleton process based on a morphological operation represented by the following formula (1) for the image signal of the bone part after the subtraction process;
Index value calculation means for calculating an index value representing the trabecular structure based on the shape information of the trabecular bone emphasized by the trabecular emphasis means;
A bone image processing apparatus comprising: a determination unit that determines a state of the bone tissue of the subject based on numerical information related to the density of the bone component and an index value representing the trabecular structure.

9. The bone image processing apparatus according to claim 8, wherein the value of n1 in the formula (1) is 2 or more. Based on two radiographic images represented by image signals of high luminance and high signal levels, each of which has a subject including a soft tissue and a bone tissue, the radiation energy distribution for energy subtraction processing being different from each other, and the bone A bone image processing apparatus for obtaining information on the density of bone components in the tissue,
  Energy subtraction processing means for performing energy subtraction processing for extracting or enhancing a desired bone portion of an image represented by the image signal based on two predetermined image signals;
  Based on the image signal of the bone part, bone density analysis means for obtaining numerical information on the density of the bone component;
  An image signal representing at least one of the two radiographic images before the energy subtraction processing, an image signal representing an added image obtained by weighting and adding the two radiographic images in correspondence with pixels, or the energy A trabecular enhancement means for emphasizing a trabecular bone using a skeleton process based on a morphological operation represented by the following formula (2) for the image signal of the bone after the subtraction processing;
  Index value calculating means for calculating an index value representing the trabecular structure based on the form information of the trabecular bone emphasized by the trabecular emphasis means;
  A bone image processing apparatus comprising: a determination unit that determines a state of a bone tissue of the subject based on numerical information relating to the density of the bone component and an index value representing the trabecular structure.

11. The bone image processing apparatus according to claim 10, wherein the value of n1 in the formula (2) is 2 or more. The bone density analyzing means is
Created from the bone mineral reference recorded in the radiographic image together with the subject, the bone mineral reference having a known bone mineral content, and the density of the bone mineral reference image appearing in the bone image and the known bone mineral content 9. The numerical information is obtained by converting a bone density in the bone image of the subject into a bone mineral density based on a density / bone mineral density conversion table. from 11 bone image processing apparatus according to any one of.

Images