JPH04176290A - Picture encoder - Google Patents

Abstract

PURPOSE:To obtain a satisfactory picture adapted to the visual sense of a man by classified into blocks based on the amount of movement detected for each block, deciding an encoding step according to the classification for each block, and encoding each block of the inputted picture. CONSTITUTION:In a moving search/correction part 9, the data of the inputted block and the data before one frame are compared and a block where the difference with the input block data becomes the minimum is retrieved so as to obtain a moving vector. In a block classification part 12, the classification of the moving vector of the moving search/correction part 9 is performed according to its distance and direction. In a quantization step control part 13, the quantization steps are finer against the block with violent movement based on the group decision result. The block with less movement is controlled rougher than the above-mentioned block. Thus, the picture on the part where a man observes can be improved.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] The present invention relates to an image encoding device used in a moving image encoding and transmitting device or a device for recording moving images on a CD ROM or the like.

[Prior Art] Generally, when transmitting image information or recording images, the amount of image information is enormous, but the line speed and cost of the line used for transmission, the capacity and capacity of the recording media, etc. From the point of view of cost etc., the amount of image information must be compressed and encoded to reduce the amount of information before being transmitted or recorded. The image signal of each small block or various prediction error signals in each small block is subjected to orthogonal transform such as discrete cosine transform, and the transform coefficients are quantized. A transform encoding device for encoding is known.

FIG. 4 shows an embodiment of a conventional transmission device using the above-mentioned transform encoding device.

An input image that is divided into blocks and sequentially sent to the input unit 1 is orthogonally transformed in an orthogonal transform unit 2 and then quantized in a quantizer 3 according to instructions from a quantization step control unit 4. The quantized result is transmitted to the outside via the multiplexer section 5 and the transmission buffer 6. At the same time, the quantized data is inversely transformed by the inverse quantization unit 7 and the inverse orthogonal transformation unit 8 to reproduce the input image, and is returned to the subtractor 10 via the motion search/compensation unit 9 to generate the next input image. The difference from the signal is taken.

[Problems to be Solved by the Invention] Generally, the region of interest in images transmitted through videophone calls, video conferences, etc. is often a human image, and the region is always accompanied by movement.

Furthermore, it has the characteristic of human vision of focusing on moving parts.

In the conventional example described above, regardless of the attribute of what the data of the input block represents, quantum processing is simply determined by whether the block is a luminance block or a color difference block, the order of the coefficients, the amount of the transmission buffer 6, etc. However, it is impossible to transmit the region of interest in the image information to be transmitted as a clearer image, and there is a problem that the points of the image to be transmitted may be lost. Ta.

Furthermore, since the coding is performed regardless of whether there is movement or not, and visual characteristics on the receiving side are not taken into account, there is a problem in that moving parts are particularly degraded and difficult to observe.

Therefore, the present invention has a simple structure, and in addition, it is possible to improve the image quality of moving parts, and also to improve the image quality of moving parts, and to improve the image quality when the area of interest to be transmitted in videophone calls, video conferences, etc. is reproduced on the receiving side. The image encoding device used in video encoding and transmission devices, etc., can encode images appropriately according to the visual characteristics and the purpose by controlling the image to be clearer than the original image. is intended to provide.

[Means for Solving the Problems] In order to achieve the above object, the present invention provides an image encoding device that divides an input image into blocks and encodes each block. a motion amount detection means; a block classification means for classifying each block based on the motion amount; an encoding step control means for determining an encoding step for each block according to the classification; The image encoding apparatus is characterized in that it has an encoding means for encoding an input image for each block.

[Operation] The motion amount detecting means detects the amount of motion for each block, and the block classification means classifies the block based on this amount of motion. In this way, the encoding step control means determines the encoding step according to the classification applied to each block, and the encoding means encodes each block of the input image according to this encoding step.

[Example] The present invention will be explained below with reference to the drawings.

FIG. 1 is a block diagram of an embodiment in which the image encoding apparatus of the present invention is applied to a moving image encoding and transmitting apparatus.

The image signal is a normal luminance signal Y and a pair of color difference signals Cb+C.
The signal r is divided into three component signals, and each component signal is sequentially input to the input section 1 in small block units. The input data of each block or various prediction error signals are subjected to orthogonal transformation such as discrete cosine transformation in orthogonal transformation section 2.
The obtained transform coefficients are quantized by a quantizer 3. At this time, the quantization step control unit 3 adaptively controls the quantization step depending on whether the input block is the luminance signal Y or the color difference signal Cb+Cr, the order of the transform coefficient, the amount of the transmission buffer 6, etc. are doing.

The quantized data is encoded in a multiplexer unit 5 suitable for transmission, stored in a transmission buffer 6, and sequentially transmitted.

On the other hand, the quantized data is inversely transformed by the inverse quantizer 7 and the inverse orthogonal transformer 8, and the transmitted image is reproduced.
In encoding a moving image, a motion search/compensation unit 9 performs interframe prediction such as motion compensation, and then a subtracter 1o performs subtraction processing with the input image signal, and the signal is input to the orthogonal transform unit 2 via a switch 11 It is designed so that If the amount of motion detected by the motion search/compensation section 9 is too large, the switch 11 is switched to the a side, and the frame is sent to the orthogonal transformation section 2 without taking a difference from the previous frame.Features of the Invention A block classification unit 12 calculates the absolute amount of motion from the motion vector from the motion search/compensation unit 9, determines a moving region of interest, and classifies each block, and a discrimination signal from the block classification unit 12. It is to have a quantization step control section 13 that varies the quantization step.

The motion search/compensation unit 9 compares the input block data with the data one frame before, searches for a block with the minimum difference from the input professional data, and obtains a motion vector.

Here, when a human being observes an image, the viewpoint is generally focused on moving parts. Such characteristics are particularly noticeable in cases where the facial expression of the other party is important, such as in a video phone call or video conference.

Another characteristic of the gaze point is that it responds more sensitively to horizontal movement than to vertical movement.

Therefore, in the block classification section 12, a second
Grouping is performed as shown in Figure (a). In FIG. 2, the horizontal axis (+X+) is the absolute amount of the motion vector in the horizontal direction, and the vertical axis (+y+) is the absolute amount of the motion vector in the vertical direction, scaled in pixel units. The grouping shown in FIG. 2(a)K is weighted in the long distance direction and in the lateral direction.

This grouping is configured to be flexible in software so that it can be grouped in an equal distance space as shown in FIG. 2(b) depending on the purpose.

As shown in FIG. 3(a), in the quantization step control unit 13, based on the group determination results, the quantization step is finely set for blocks with a lot of movement, and the quantization step is set finely for blocks with little movement. Control the transformation step more coarsely than for fast moving blocks. FIG. 3(a) shows an example of controlling the quantization steno amplifier proportional to the amount of movement. In addition, in the same figure, the quantization step width N on the vertical axis
represents the initial quantization step width.

In this simple way, by adaptively controlling the quantization step according to the amount of motion of each input block, it is possible to improve the image quality of the part that the human is looking at. This makes it possible to provide an image encoding device adapted to visual characteristics.

Quantization step performed by the quantization step control section 13.

Another example of step determination is shown in FIG. 3(b).

For example, when considering video telephones and video conferences, an important part of the transmitted information is facial expressions and the like.

If the image to be transmitted is of one or two people, the amount of movement of each part of the face that expresses facial expressions is approximately 10 pixels. Therefore, the control in the quantization step control section 13 is performed as shown in FIG. 3(b).
By finely controlling the quantization step for the movement before and after pixels, it is possible to improve the image quality of the region of interest, especially the face, and to obtain detailed information about the center of the image. It is possible to provide an image encoding device that can obtain the following.

〔Effect of the invention〕

As is clear from the above, by controlling the quantization step according to the amount of motion, the image encoding device according to the present invention can produce a high-quality image, especially for the part to be watched, and is compatible with human visual characteristics. It is possible to obtain a suitable and good image.

Furthermore, by using the image encoding device according to the present invention as a moving image encoding and transmitting device and adapting it to video telephone calls and video conferences, facial expressions, etc. can be transmitted in detail, resulting in efficient encoding adapted to the purpose of transmission. can do.

[Brief explanation of drawings]

FIG. 1 is a block diagram showing an embodiment of an image encoding device according to the present invention, FIG. 2 is a diagram explaining a method of classifying input image blocks by a motion amount determining section, and FIG. 3 is a block diagram showing an example of a quantization step control section. A diagram explaining how to determine the quantization step,
FIG. 4 is a block diagram of a conventional image encoding device. , N L g'j, 'i., Input section, 2.. Orthogonal transformation section,
3. Quantization section, 7. Inverse quantization section, 8.. Inverse orthogonal transformation section, 9.. Motion search fist compensation section, 10.. Subtractor, 11.
...Switch, 12.. Block classification section, 13.. Quantization step control section. Agent Patent Attorney Ume 1) Katsu (and 2 others) Figure 4

Claims (1)

[Claims] 1. An image encoding device that divides an input image into blocks and encodes each block, comprising: a motion amount detection means for detecting the amount of motion of each block; a block classification means for classifying the input image; an encoding step control means for determining an encoding step for each block according to the classification; and an encoding means for encoding the input image for each block according to the encoding step. An image encoding device comprising: