JPS5826073B2 - Adaptive sampling processing method for drawn images - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention provides an adaptive sampling processing method for drawn images, particularly a system for inputting, transmitting, and displaying drawn images such as handwritten characters and figures. ,
Adaptively determine the sample points to be transmitted in the drawn image,
The present invention relates to an adaptive sampling processing method for extracted drawn images.

Conventionally, in a system that transmits a drawn object through a narrowband communication line such as a telephone line, there is a method of sampling position coordinate signals obtained from a handwritten figure input device such as a tablet, and a method of sampling position coordinates at fixed time intervals. , a method has been used in which a fixed window is provided around one sampling point, the coordinates of the window are detected, and that point is set as the next sampling point.

In the former case, the difference between the sampled points, that is, the encoding is performed in relative coordinate format, and in the latter case, the directional component of the points across the window is encoded.

In such a method, although it is possible to transmit data using a narrowband transmission path for normal writing speed conditions, 1. When inputting a drawing at high speed, the former method has drawbacks such as deterioration in display quality, such as curves being restored as broken lines due to a lack of sample points, and line lengths being shortened.

In addition, since the latter method increases the amount of code to be transmitted,
It had drawbacks such as causing transmission delays.

In order to solve these drawbacks, the present invention extracts the sample points to be transmitted using a mask that depends on the drawing input speed (elapsed time) and direction, and eliminates geometric distortion of the reproduced image. The purpose is to keep it small and reduce the number of sample points, and the drawings will be explained in detail below.

FIG. 1 shows the configuration of an embodiment of the present invention, where a represents a transmitting end and b represents a receiving end.

Further, 1 is a graphic tufflet for inputting handwritten figures, etc., 2 is a sampling input processing circuit, 3 is an encoder,
4 is a transmission path, 5 is a decoder, 6 is a display processing circuit, and 7 is a monitor.

To operate this device, when a character or a figure is written on the graphic tablet 1 at the transmitting end a, a coordinate signal corresponding to the input figure is output from the graphic tablet 1.

The sampling input processing circuit 2 performs sampling processing to delete coordinate signals unnecessary for restoring the input figure by a method described later.

The sampled coordinate signal is encoded by an encoder 3 according to an encoding method convenient for transmission, and output to a transmission line 4.

At the receiving end, the received signal is decoded by a decoder 5 to obtain a coordinate signal corresponding to the input point sequence.

The display processing circuit 6 performs interpolation on the point sequence decoded by the decoder 5 by connecting the points with straight lines, and outputs it to the monitor 7 as a video signal.

Next, sampling by the sampling input processing circuit 2 will be described using FIG. 2 and subsequent figures.

The types of masks used to extract the points to be transmitted from the points sampled with sufficient density and high speed are, for example, the mask ML for straight parts shown in Figure 2 and the mask ML for curved parts shown in Figure 3. There are two types of mask MB.

Figures 2 and 3 (where C is the input curve for drawing, P
is one sample point to be transmitted, and P' is the sample point to be transmitted next after point P.

The mask ML for the straight line portion is used when the writing speed is fast.

If the writing speed is fast (in this case, the variation in the drawing input direction is relatively small, and detailed parts are not input, the error when approximating with a straight line is within the tolerance value W, as shown in Figure 2) A long and narrow rectangular mask is determined along the drawing input direction near the point Pft.

In such a case, the next sample point to be transmitted is the first sample point P' that exits this mask.

However, even when extracting sample points using the straight line mask ML, there may be cases where the drawing input speed suddenly slows down or the input direction suddenly changes by nearly 180 degrees. Even if P is inside the straight line mask ML, in the former case, a certain period of time td has elapsed since the point P was sampled, and in the latter case, the point where the input direction suddenly changed is the next point, as shown in Figure 4. Let P' be a sample point to be transmitted.

The mask MB for the bend shown in FIG. 3 is used when the writing speed becomes slow, that is, when the sample point does not go outside the mask (ML or MB) even after a certain time td as shown in FIG.
Or it is used when the input direction suddenly changes.

This mask extracts the next sample point to be transmitted when the distance between the sample points exceeds a certain value, regardless of the input direction, and efficiently samples small changes.

FIG. 5 shows an explanatory diagram showing the above explanation in the form of a flowchart.

As you can see from the figure, first check with the mask MB,
Sample points are extracted using the mask ML as much as possible.

In addition, Figure 6A schematically shows the state of sampling.
P l r meeting, + +2 t P29◆3. ...
... is a sequence of points sampled at high speed at time interval t3, among which Pl, P2 . P3... is a sequence of points to be transmitted.

Further, P3 is set as a sample point to be transmitted because the time limit td has elapsed.

Moreover, FIG. 6B shows this time relationship.

As described above, the drawn image can be sampled adaptively by sampling the drawn image at high speed and adaptively changing the mask in order to extract points to be transmitted from the point sequence.

Although the shape of the mask ML for the straight portion is rectangular, it does not need to be rectangular, and any other shape may be used as long as the distortion can be kept within an allowable range.

In addition, the mask MB for the bent part is also shown as a square in the figure, but
For example, there is no problem even if it is yen.

Furthermore, the number of types of masks to be used does not have to be two, and the number of types may be increased.

The criteria for setting the orientation of the mask ML for the straight portion are, taking the case of P2 shown in FIG. There are methods such as taking minutes as the standard.

Further, the resolution of orientation does not need to be so strict, and processing can be performed by classifying the orientation into approximately 8 to 16 predetermined orientations as necessary.

As explained above, according to the present invention, it is possible to determine sample points to be transmitted through efficient real-time sequential processing so that distortion is reduced when linear interpolation is performed on the display side. Therefore, when drawing is input at high speed, the deterioration of display quality is minimized, and when input is at low speed, unnecessary information is suppressed, and limited transmission is possible without sacrificing minimum resolution. This has the advantage of being able to efficiently transmit drawn image signals at high speed.

Furthermore, even when a drawn image signal is stored, there is an advantage that the amount of information can be compressed to the same extent as when it is processed as a still image.

[Brief explanation of the drawing]

FIG. 1 is a schematic configuration diagram showing an embodiment of the present invention, FIG.
3 and 4 are explanatory diagrams for explaining a method of extracting sample points to be transmitted using a mask, FIG. A schematic diagram showing how points are extracted is shown. DESCRIPTION OF SYMBOLS 1...Tablet, 2...Sampling input processing circuit, 3...Encoder, 4...Transmission path, 5...Decoding device, 6...display processing circuit, 7...monitor, C...drawing input curve, P...one sample point to be transmitted, Pl...・
...Sample point to be transmitted next to P, ML...Mask for straight section, MB...Mask for curved section, Pl,
P2. P3゜P4... Point sequence to be transmitted /)1
．． 'Meeting 22. ...The point where the sample was sampled at high speed but it was not transmitted.

Claims (1)

[Claims] 1. A means for detecting the position coordinates of a human power point, a means for sampling the detected coordinates as a discrete point sequence at a predetermined time interval with a density within a predetermined range, and a means for determining the points to be transmitted from the sampled point sequence. In a drawing image sampling processing method that has means for extracting, a plurality of masks with different shapes are provided, and when extracting a point to be transmitted from a sampled point sequence, one transmitted sample point is Correspondingly, a predetermined one of the plurality of masks is associated with the pen speed and direction of the transmitted sample point, and the sample point selected based on the mask is transferred to the next sample point (the sample to be transmitted). An adaptive sampling processing method for a drawn image is characterized in that the sampling points to be extracted as points and sequentially transmitted are determined.