JP7078255B2 - Ear print database generation system - Google Patents

Description

The present invention relates to a system for generating a database of ear prints.

Focusing on the fact that there are significant individual differences in the uneven shape of the human auricle, many studies on personal authentication using images of the auricle have been conducted (for example, Non-Patent Document 1).

On the other hand, in some countries such as the Netherlands and the United Kingdom, ear prints left on the crime scene (traces of the criminal pressing the pinna against the window or wall) have been used for criminal investigations. A photo of the suspect's pinna is collated with the remaining residual ear print.

However, Non-Patent Document 2 reports a research result that the authentication rate is only 50% in the matching by pattern matching of the ear print image and the auricle photograph, which is not practical as an authentication method.

Katsuyuki Shinohara, "Outline features of auricular cartilage in personal authentication using auricle", Proceedings of the Society of Imaging Electronics, 2006. Nobuyuki Nishiuchi, Lucas Bitswaff, et al., "Study on Similarity Measurement between Auricular Image and Residual Earprints", Academic Techniques, 117 (165), pp.43-45, 2017.

The present invention has been made in view of the above, and an object of the present invention is to provide a system for generating a database of ear prints applicable to criminal investigation.

The inventor was inspired to generate a database of ear prints intended to be collated with residual ear prints at crime scenes. However, since the pattern of the ear print changes greatly depending on the amount of force that presses the pinna, it is expected that the ear print remaining at the crime scene will exhibit various patterns. As a result of diligent studies in consideration of this point, the present inventor came up with the following configuration and came up with the present invention.

That is, according to the present invention, a square columnar prism having a trapezoidal cross section, the first surface corresponding to the upper bottom of the trapezoid, the second surface corresponding to the lower bottom of the trapezoid, and the like. A prism having a third surface corresponding to the slanted side of the trapezoid, a light source that irradiates the illumination light from below the second surface toward the first surface, and the first surface are pressed against the prism. A high-speed camera arranged so as to capture the light reflected from the second surface and emitted from the third surface among the light reflected on the first surface, and the ear canal on the first surface. A load detecting means for detecting a load applied to the prism by being pressed against the prism, a captured image recording means for recording an image captured by the high-speed camera in association with the imaging time, and the load detecting means. The load recording means that records the load detected by the light in association with the detection time, the mapping means that associates the captured image taken at the time when the predetermined load is detected with the load, and the predetermined load. An ear print image generation means for generating an ear print image by directly facing the captured image, and an ear print image storage means for associating the generated ear print image with a load corresponding to the ear print image and storing the generated ear print image in a database. And, including the ear print database generation system is provided.

As described above, the present invention provides a system for generating a database of ear prints applicable to criminal investigations.

The figure which shows the ear print database generation system of this embodiment. The figure which shows the ear print database generation system of this embodiment. The figure which shows the image taken by the high-speed camera. The flowchart of the process executed by the computer of this embodiment. The figure which shows the table generated by the computer of this embodiment. The figure which shows the ear print image accumulated in the ear print database of this embodiment.

Hereinafter, the present invention will be described with reference to the embodiments shown in the drawings, but the present invention is not limited to the embodiments shown in the drawings. In each of the figures referred to below, the same reference numerals are used for common elements, and the description thereof will be omitted as appropriate.

FIG. 1 shows the configuration of the ear print database generation system 100 according to the embodiment of the present invention. The ear print database generation system 100 of the present embodiment includes a prism 10, a load cell 20, a light source 30, a pedestal 40, a high-speed camera 50, and a computer 60.

The prism 10 is a square columnar transparent body having a right-angled trapezoidal cross section, and is fixed on the pedestal 40 via a load cell 20. The prism 10 has a rectangular surface 10a corresponding to the upper base of the right trapezoid, a rectangular surface 10b corresponding to the lower base of the right trapezoid, and a rectangular surface 10c corresponding to the diagonal side of the right trapezoid, and the rectangular surface 10b and the rectangle. The angle formed by the surface 10c is 45 °. Then, in the present embodiment, when collecting the ear print, the subject is instructed to press the auricle against the rectangular surface 10a.

The load cell 20 is a load detecting means for detecting a load applied to the prism 10 when the auricle of the subject is pressed against the rectangular surface 10a, and the detected value is transmitted to the computer 60 in real time at an appropriate sampling rate. It is configured.

The light source 30 is a light source for irradiating the target's ear canal pressed against the rectangular surface 10a with illumination light, is arranged in the center of the load cell 20, and is rectangular from below the rectangular surface 10b of the prism 10. Illumination light is emitted toward the surface 10a. As the light source 30, it is preferable to use an LED light source having high straightness.

As shown in FIG. 2, the illumination light emitted from the light source 30 is incident from the rectangular surface 10b, then passes through the inside of the prism 10 in the height direction of a right-angled trapezoid, and is emitted from the rectangular surface 10a. It reaches the ear canal that comes into contact with the light source and diffusely reflects. Then, a part of the light reflected on the pinna is reflected on the rectangular surface 10b inside the prism 10 and emitted from the rectangular surface 10c.

Here, the high-speed camera 50 is positioned at an appropriate position where the reflected light emitted from the rectangular surface 10c can be captured and at an appropriate position where the light source 30 is not reflected, and the subject's ear canal is rectangular. While pressed against the surface 10a, it is configured to continuously capture images at an appropriate frame rate (eg, hundreds of frames per second) and transmit the captured images to the computer 60 in real time. In this embodiment, it is desirable that the sampling rate of the load cell 20 is higher than the frame rate of the high-speed camera 50.

FIG. 3 shows a captured image transmitted by the high-speed camera 50 to the computer 60. As shown in FIG. 3, in the present embodiment, a portion (that is, an ear print) in contact with the rectangular surface 10a of the auricle of the subject is acquired as a bright image. However, the image acquired at this time is not a face-to-face image in which the ear print is viewed from the front, but an inclined image in which the ear print is viewed from an oblique direction.

Here, returning to FIG. 1, the functional configuration of the computer 60 will be described.

The computer 60 includes a captured image recording unit 62, a load recording unit 63, an association unit 64, an ear print image generation unit 65, an ear print image storage unit 66, and a database 68.

The captured image recording unit 62 is a means for recording the captured image captured by the high-speed camera 50 in association with the imaging time while the subject's auricle is pressed against the rectangular surface 10a of the prism 10. ..

The load recording unit 63 is a means for recording the load detected by the load cell 20 in association with the detection time while the auricle of the subject is pressed against the rectangular surface 10a.

The association unit 64 is a means for associating a captured image captured at a time when a predetermined load is detected with the load.

The ear print image generation unit 65 is a means for generating an ear print image by facing the captured image associated with a predetermined load.

The ear print image storage unit 66 is a means for associating the generated ear print image with the load corresponding to the ear print image and storing the generated ear print image in the database 68.

The functional configuration of the computer 60 has been described above, and subsequently, the processing executed by the computer 60 will be described with reference to the flowchart shown in FIG.

In the present embodiment, as described above, while the subject's auricle is pressed against the rectangular surface 10a, the high-speed camera 50 transmits the captured image to the computer 60 in real time, and the load cell 20 is sent to the computer 60. Send the load in real time. In response to this, the captured image recording unit 62 and the load recording unit 63 execute step 101 and step 102 in parallel.

In step 101, the captured image recording unit 62 associates the captured image received from the high-speed camera 50 with the captured time while the auricle of the subject is pressed against the rectangular surface 10a in FIG. 5 (a). ) Is recorded in the correspondence table 500. Here, the correspondence table 500 includes a field 501 for storing a captured image and a field 502 for storing a captured time, and in step 101, the captured image received from the high-speed camera 50 is stored in the field 501. The reception time of the captured image (system time of the computer 60) is regarded as the imaging time and stored in the field 502. In another embodiment, the high-speed camera 50 may be made to transmit the imaging time of the captured image and store it in the field 502.

On the other hand, in step 102, the load recording unit 63 associates the load received from the load cell 20 with the detection time while the auricle of the subject is pressed against the rectangular surface 10a, and is shown in FIG. 5 (b). It is recorded in the corresponding correspondence table 600 shown. Here, the correspondence table 600 includes a field 601 for storing the load and a field 602 for storing the detection time, and in step 102, the load received from the load cell 20 is stored in the field 601. The reception time of the load (system time of the computer 60) is regarded as the detection time and stored in the field 602. In another embodiment, the load cell 20 may be made to transmit the load detection time and store it in the field 602.

The above-mentioned steps 101 and 102 are executed until the detected value of the load received from the load cell 20 reaches the predetermined target value A (step 103, No), and when the detected value of the load reaches the target value A (step 103, No). Step 103, Yes), the process proceeds to step 104.

In the following step 104, the mapping unit 64 was photographed at a time when a predetermined load was detected based on the mapping table 500 (FIG. 5 (a)) and the mapping table 600 (FIG. 5 (b)). Attach the captured image to the load. Here, the predetermined load means a discrete value of the load.

Hereinafter, values from 0 to the target value A are carved at intervals of 0.1 [N] (0.1 [N], 0.2 [N], 0.3 [N], ... Target value A [N]). The process executed in step 104 will be described by taking the case where is set to a predetermined load as an example.

First, referring to the correspondence table 600 (FIG. 5 (b)), a record in which 0.1 [N] is stored in the field 601 is specified, and the detection time stored in the field 602 of the record is extracted. .. If the record in which 0.1 [N] is stored does not exist, the detection time stored in the field 602 of the record in which the value closest to 0.1 [N] is stored is extracted. Here, when a plurality of detection times are extracted, one detection time is extracted from the detection times according to a predetermined rule (for example, the oldest time is extracted).

Next, with reference to the correspondence table 500 (FIG. 5A), the record in which the extracted detection time is stored in the field 502 is specified, and the captured image stored in the field 501 of the record is extracted. If the record in which the detection time is stored does not exist, the captured image stored in the field 501 of the record in which the time closest to the detection time is stored is extracted.

Finally, the extracted captured image is associated with 0.1 [N] and recorded in the table. After that, the remaining load (0.2 [N], 0.3 [N], ... Target value A [N]) is processed in the same manner.

FIG. 5C shows the mapping table 700 generated in step 104. The correspondence table 700 includes a field 701 for storing a load, a field 702 for storing a captured image, and a field 703 for storing an ear print image, and when step 104 is completed, the correspondence table 700 includes a field 701. Field 703 is left blank.

In a subsequent step 105, the ear print image generation unit 65 generates an ear print image by facing the captured images stored in the field 702 of each record of the correspondence table 700. This is a process based on the fact that the captured image recorded in the previous step 101 is an inclined image when the ear print is viewed from an oblique direction, and the captured image stored in the field 702 of each record is subjected to projective conversion. Then, the ear print is converted into a face-to-face image that would be obtained when the ear print is imaged from the front, and the face-to-face image is stored in the field 703 of the corresponding record as the ear print image. FIG. 5D shows a mapping table 700 at the time when step 105 is completed.

Finally, in step 106, the ear print image storage unit 66 associates the ear print image generated in the previous step 105 with the corresponding load and stores it in the database 68. Specifically, the ear print image stored in the field 703 of the correspondence table 700 (FIG. 5 (d)) and the load stored in the field 701 are associated and stored in the database 68.

As described above, according to the present embodiment, for one pinna, as conceptually shown in FIG. 6, a plurality of ear print images are linked to the load and created in a database. As described above, the ear prints remaining on the crime scene are expected to exhibit various patterns, but according to the present embodiment, the ear print patterns that may remain on the crime scene are comprehensive. Since it is registered in, it is highly practical in criminal investigation.

Although the present invention has been described above with embodiments, the present invention is not limited to the above-described embodiments, and the actions and effects of the present invention can be achieved within the scope of other embodiments that can be inferred by those skilled in the art. As long as it works, it is included in the scope of the present invention.

10 ... Prism 10a ... Rectangular surface 10b ... Rectangular surface 10c ... Rectangular surface 20 ... Load cell 30 ... Light source 40 ... Pedestal 50 ... High-speed camera 60 ... Computer 62 ... Image capture image recording unit 63 ... Load recording unit 64 ... Corresponding unit 65 ... Ear print image generation unit 66 ... Ear print image storage unit 68 ... Database 100 ... Ear print database generation system 500 ... Correspondence table 501, 502 ... Field 600 ... Correlation table 601, 602 ... Field 700 ... Correspondence table 701, 702 , 703 ... Field

Claims (3)

A square columnar prism having a trapezoidal cross section, the first surface corresponding to the upper base of the trapezoid, the second surface corresponding to the lower base of the trapezoid, and the third surface corresponding to the oblique side of the trapezoid. With a prism, which has a surface of
A light source that irradiates the illumination light from below the second surface toward the first surface,
A high-speed camera arranged so as to capture the light reflected from the second surface and emitted from the third surface among the light reflected by the auricle pressed against the first surface. ,
A load detecting means for detecting a load applied to the prism by pressing the pinna against the first surface, and a load detecting means.
An image pickup image recording means for recording an image captured by the high-speed camera in association with the image pickup time.
A load recording means that records the load detected by the load detecting means in association with the detection time, and a load recording means.
Associating means for associating a photographed image taken at a time when a predetermined load is detected with the load, and
An ear print image generation means for generating an ear print image by directly facing an captured image associated with a predetermined load,
The ear print image storage means that stores the generated ear print image and the load corresponding to the ear print image in the database in association with each other.
Ear print database generation system including. The light source is an LED light source.
The ear print database generation system according to claim 1. The prism has a right-angled trapezoidal cross section, and the angle between the second surface and the third surface is 45 °.
The ear print database generation system according to claim 1 or 2.

Images