KR0153675B1 - A encoder with region discriminator - Google Patents

Abstract

The present invention relates to an encoder having an area discriminator. The encoder according to the present invention divides an image of all frames and compares the image of every frame with each segment, and then selects an area by determining whether it is a still image area or a moving image area. There is an advantage in that the area discrimination can be made easy by adding a part of the same.

Description

Encoder with area discriminator

1 is a diagram illustrating a sub-sampling method of HD-MAC, (a) is a sub-sampling diagram for each mode of the HD-MAC luminance signal, and (b) is a two-dimensional filter for preventing overlapping by mode according to the HD-MAC luminance signal sample pattern Passband display diagram.

2 is a diagram illustrating a sub-sampling method of MUSE, (a) is a display diagram showing a field offset sub-sampling method of MUSE, (b) is a flowchart of a luminance signal subsampling system of MUSE, and (c) a subsampling schematic diagram of MUSE.

(A) of FIG. 3 shows the entire pixel signal of the original image, (b) shows the first sub-sampling of the image signal of (a), (c) shows the first sub-sample of (b) (2) is a diagram showing a passband of a filter having a rhombic characteristic used for subsampling.

4 is a partial block diagram of an encoder according to the present invention.

* Explanation of symbols for main parts of the drawings

10 ... Still image area subsampling section 12 ... Pixel selector

14 ... moving picture area subsampling unit 16 ... first frame memory

18 ... second frame memory 20 ... video splitter

22 ... Area discriminator 24 ... Area selector

The present invention relates to an encoder equipped with an area discriminator. In particular, an image segmentation is performed on an image of the previous frame and compared with the image of the previous frame for each segment. The present invention relates to an encoder for facilitating area discrimination.

HDTV (High Definition Television) has more than twice the vertical and horizontal resolution of current televisions and not only improves the problems of current televisions such as cross color, but also has an aspect ratio of 16: 9 and achieves CD sound quality. With digital audio performance, we say next-generation TV which provides clearer picture and clearer sound quality than current TV.

HDTV systems can be classified into two types based on the transmission method.

The first is a sampled value transmission method in which a digital signal is regarded as an analog signal and transmitted, such as Japan's MUSE (Multiple Sub-Nyquist Sampling Encoding) and Europe's HD-MAC. The second is the digital-digital method, which is digitalized from signal processing to transmission, and is now rapidly reaching the stage of practical application, starting with the United States.

Japan's MUSE and Europe's HD-MAC compress video bands by subsampling and time division multiplexing (TDM) and transmit them in the form of analog signals. On the other hand, in the full-digital scheme, as long as the signal-to-noise ratio of the receiver remains above a certain threshold, even if a channel error occurs, error correction is performed so that the quality of the transmitter is almost completely transmitted to the receiver.

Because HDTV signals have a huge amount of data, the need arises to compress the data for transmission. Although there are various compression methods, a band coding method, which has recently been of interest, is divided into a plurality of consecutive bands by using a plurality of filters in a frequency band, and subsampling for each signal series ( Subsampling), encoding, and transmission method.

The sub-sampling method is based on HD-MAC in Europe and MUSE in Japan among HDTV systems.

FIG. 1 is a diagram of a sub-sampling scheme of HD-MAC. There are three different spatial sub-sampling schemes in HD-MAC. That is, there are subsampling methods used in 80ms mode, 40ms mode, and 20ms mode, respectively. The refresh interval is 80ms in the stop mode (speed range: 0-0.5 sample / frame), the refresh interval is 40ms in the slow movement mode (0.5-12 sample / frame), and the refresh interval in the move mode (more than 12 samples / frame). Is 20ms.

(A) of FIG. 1 is a sub-sample of the luminance signal for each mode in the high-quality sample grid, and the displayed numerical values are sampled values of the fields. (The size of the block is 16 * 16, but here only 8 * 8 is shown.)

In 80ms mode, samples of 1, 2, 3, 4 and numbered positions are transmitted for 80ms (4 fields). In the 80ms mode, it is assumed that there is no motion for 80ms in the 80ms mode, so samples may be taken in only two fields (one frame) out of four fields (two frames). However, when transmitting a video signal to D / D2-MAC, only 8 * 4 samples (32 samples) can be sent per block, so samples in one position are second, samples in positions 2, 3, and 4 in the first field, respectively. Send in the third and fourth fields. Here, 32 samples must belong to 8 * 4 blocks, so the samples of lines 3, 7, 11, and 15 are sandwiched between the samples of lines 1, 5, 9, and 13, and the samples of lines 4, 6, 10, and 14, respectively. The data is sandwiched between the two transmissions and called shuffling. (The first line is defined as line 1.) In the decoder, the samples collected during 4 fields (128 samples) are collected and interpolated into 16 * 16 blocks.

In 40ms mode, samples of positions 1 and 2 are written for 40ms (2 fields). In FIG. 1A, since the odd-numbered lines belong to the odd field and the even lines belong to the even field, only the samples of the odd field are transmitted in the 40ms mode. However, when transmitting a video signal to D / D2-MAC, only 8 * 4 samples (32 samples) can be sent per block, so samples at one position are placed in the odd field and samples at the two positions are placed in the even field. Send it.

Here, the samples are sampled in an 8 * 4 structure, so there is no need for shuffling. In the decoder, samples (64 samples) sent during two fields are collected and interpolated into 16 * 8 blocks to form a 40ms mode block in the lake field, and 40ms blocks of even fields are filled by a motion compensation method.

In the 20ms mode, one field is transmitted for 20 ms (one field), and one position is transmitted in an even field, and two positions are transmitted in an even field. In this case, the samples are sampled in a 4 * 8 structure, which requires shuffling in shape. In the decoder, samples (32 samples) sent during one field are collected and interpolated into 16 * 8 blocks to form 20ms mode blocks in odd or even fields, respectively.

The subsampling of each mode is spatial deduction, and two-dimensional low pass filtering should be performed to prevent aliasing.

(B) of FIG. 1 shows the passband of the two-dimensional low-pass filter for each mode. Since all three modes of subsampling are basically sampled sampling, the passband is shown in a rhombus shape. In (b), only the upper right side of the lozenge is shown.

Even in the case of diagonal sampling (Quincunx Sampling), a rectangular low pass filter can realize a filter that prevents aliasing.However, since the human eye is sensitive to vertical or horizontal high frequency components rather than diagonal, the vertical or horizontal Use a lozenge filter with high frequency emphasis. However, when the rhombic filter is used, the vertical and horizontal components are not separated into two one-dimensional filterings, and thus only two-dimensional filtering is possible.

The sub-sampling method of MUSE has the same sampling pattern for each mode of moving picture and still picture, which is different from the method used in HD-MAC. There is no need to transmit, and in particular, there is an advantage in that a serious image quality deterioration due to an error in mode information between the encoder and the decoder can be excluded.

(A) of FIG. 2 shows a field offset subsampling method of MUSE, in which the sample position is changed for each field, and the interfield offset subsampling takes a sampling pixel at 4 ms field sequential. In addition, since the image quality deteriorates, the correlation between adjacent scan lines, fields, and frames is strong, and almost the same shape of the picture or pattern, so interpolation is used to compensate for insufficient sample values.

(B) of FIG. 2 shows a flow chart of the luminance (Y) signal subsampling system of MUSE. The luminance (Y) signal divides the signal band into a still picture area and a moving picture area, and a still picture requires a fine picture. When taking a wider band and extracting a sample value for each pixel, one pixel is staggered in the horizontal direction between fields, that is, an inter-field offset sampling is performed on sample values between two adjacent fields, and a moving image area does not need a wide band. The sample values in the field are added to the still picture region through the prefilter. The RY and BY signals, which are the color difference signals (C), are also extracted for each scan line, and the first line is the RY signal, the second line is the BY signal, and the third line is the sequential signal. Perform offset subsampling.

(C) of FIG. 2 is a subsampling schematic diagram of the MUSE, wherein the time compression of the input unit performs compression of 11/12 for simple integer conversion of the sampling frequency of the studio standard 74.25 MHz and the input sampling frequency of the MUSE encoder 48.6 MHz. . The final subsampling signal of the chrominance signal is subjected to time compression of 1/4 to perform time multiplexing on the blanking interval of the luminance signal by equalizing the luminance signal and the sampling rate.

The image segmentation used in the present invention will now be described.

Image segmentation refers to a process of dividing an image into components or objects. Often, the purpose of image segmentation is to identify the boundary or boundary of the physical object in the three-dimensional space. The basic assumptions in image segmentation are that the surface of an object is not bumpy and homogeneous, so the contrast of the image should change smoothly, and the contrast should change sharply at the boundary. Most of these assumptions are made, but not always. In cases where similar surfaces are adjacent to each other, the boundary line may not appear clearly, or various patterns or letters may be written on the surface, and surface texture, that is, texture or noise, may be a problem. Sometimes it is difficult.

Image segmentation can be achieved by two main ideas: discontinuity in contrast and similarity in contrast. Focusing on the discontinuity of the contrast, the key is to find the discontinuity, and based on the similarity of the contrast, the point is to continuously find pixels with similar contrast and classify them as being in the same area.

Looking at the image segmentation method, there are methods such as isolated search, line search, and contour detection, which focus on discontinuity, and region segmentation and region segmentation on the image segmentation method. And region splitting and merging, or thresholding.

(A) of FIG. 3 shows the entire pixel signal of the original image, (b) is the primary sub-sampled for (a), and (c) is the secondary sub-sample for (b) primary sub-sampled. Sampled.

Subsampling can be said to take out pixels in the middle without using all the pixels of the original image in the shape of a checkerboard. For example, subtracting one pixel in a diagonal direction from (a) of FIG. This is referred to as quincunx subsampling. Quincunx Subsampling is used a lot in image processing because of the least noticeable noise that occurs when sampling. Even in the case of Quinxunx Subsampling, a rectangular low pass filter can realize a filter that prevents aliasing.However, since the human eye is sensitive to high frequency components in the vertical or horizontal direction rather than the diagonal direction, Use a rhombic filter with high frequency emphasis. However, when a rhombic filter is used, the vertical and horizontal components are not separated into two one-dimensional filterings, and thus only two-dimensional filtering can be implemented.

(D) in FIG. 3 is a filter having a rhombic shape for subsampling, and after filtering the original image as shown in (a), one pixel point position ( After inserting the pixel of O mark), perform the rhombic filtering as shown in (d) again. Sub-sampling an image such as the first sub-sampled (b) secondly can obtain an image as shown in (c) of FIG. 3.

Since the human visual structure does not recognize the resolution of the subject well in the moving image area in which the subject moves, and the resolution of the subject is well recognized in the still image area in which the subject is stationary, the resolution of the still image area is increased. There is a need.

Therefore, since the resolution must be increased in the still image area of the subject, only the pixels of one pixel point position (O display) are transmitted in the image of FIG. 3B, and the pixels of the two pixel point position (△ display) are the pixels of the previous frame. Send them using That is, in the still image area, pixels at the one pixel point position (○ mark) and pixels at the two pixel point position (Δ mark) are alternately selected for each frame and transmitted. In the moving image area of the subject, as shown in (c) of FIG. 3, the pixels of one pixel point position (O display) subsampled twice are transmitted.

However, there is a problem in the conventional encoder that a complicated method is used to discriminate whether it is pixel information of a still picture area or pixel information of a moving picture area.

Accordingly, in order to solve the above-mentioned problems, the present invention divides the image of the previous frame and compares the image with the segment of the previous frame for each segment. The purpose is to facilitate region discrimination by adding.

The encoder of the present invention for achieving the above object comprises: a still picture region subsampling section for subsampling pixels corresponding to a still picture region of an input image signal; A pixel selector for alternately selecting and transmitting the pixels subsampled by the still image area subsampling unit for each frame; A moving image region subsampling unit for subsampling pixels corresponding to the moving image region of the input image signal twice; A first frame memory for storing image information of previous frames; A second frame memory for storing image information of all previous frames; An image divider for dividing an image to determine an image boundary line or boundary surface of all frames stored in the first frame memory; An area discriminator for discriminating an area by comparing image information of all previous frames stored in the second frame memory with image information of all frames divided by the image divider for each segment; And an area selector for selecting image information of a still picture area or a moving picture according to the area determined by the area discriminator.

Hereinafter, with reference to the accompanying drawings will be described in detail with respect to the present invention.

4 is a block diagram showing a partial configuration of an encoder according to the present invention, comprising: a still image region subsampling section 10 for subsampling pixels corresponding to a still image region of an input image signal; A pixel selector 12 for alternately selecting and transmitting the pixels subsampled by the still image area subsampling unit per frame; A moving image area subsampling unit 14 for subsampling pixels corresponding to the moving image area of the input image signal twice; A first frame memory 16 for storing image information of previous frames; A second frame memory 18 for storing image information of all previous frames; An image divider (20) for dividing an image to grasp an image boundary line or boundary surface of all frames stored in the first frame memory (16); An area discriminator (22) for comparing the image information of all the previous frames stored in the second frame memory (18) with the image information of the previous frame divided by the image divider (20) by segment; And an area selector 24 for selecting image information of one of the still picture area and the moving picture area according to the area determined by the area discriminator 22.

Referring to FIGS. 4 and 3, the operation of the present invention is as follows.

The still image area subsampling unit 10 is a part for subsampling pixels corresponding to the still image area among the input image signals. When subsampling the pixels of the still image area, an image as shown in FIG. 3B can be obtained. .

In the pixel selector 12, the pixels subsampled by the still image area subsampling unit 10 are alternately transmitted for each frame. At this time, pixels of one pixel point position (O display) or two pixel point position (△ display) are transmitted. Pixels at one pixel point position among the pixels of the second pixel are selected and transmitted, and pixels at the two pixel point position (Δ) are transmitted using the pixels of the previous frame.

The moving image area subsampling unit 14 subsamples the pixels corresponding to the moving image area of the input image signal twice and transmits the pixels shown in FIG.

The first frame memory 16 stores image information of all frames as a reference for discriminating whether it is a still picture area or a moving picture area.

The second frame memory 18 stores image information of all previous frames for comparison with previous frame image information of the first frame memory 16.

The image splitter 20 divides the image into segments in order to determine an image boundary line or boundary surface of all frames stored in the first frame memory 16.

The area discriminator 22 compares the image information of all the frames stored in the second frame memory 18 with the image information of all the frames divided by the image divider 20 for each segment to determine an area. .

The area selector 24 serves to select image information of one of the still picture area and the moving picture area according to the area determined by the area discriminator 22.

As described above, the present aspect divides the image of the previous frame by image, compares it with the image of the previous frame by segment, and discriminates whether it is a still image area or a moving image area, and then adds a portion for selecting an area to the encoder. It can facilitate the effect.

Claims (1)

A still picture region subsampling section 10 for subsampling images corresponding to the still picture region of the input image signal; A pixel selector 12 for alternately selecting and transmitting the pixels subsampled by the still image area subsampling unit per frame; A moving image area subsampling unit (14) for subsampling pixels corresponding to the moving image area of the input image signal over a second order; An encoder comprising a region selector 24 for selecting pixel information of one region from among subsampled pixel information of the still picture region selected by the pixel selector 12 and pixel information of the subsampled moving image region over a second order; A first frame memory (16) for receiving the input image signal and storing image information of all frames; A second frame memory (18) for storing image information of all previous frames from the first frame memory (16); An image divider 20 for dividing an image into segments in order to identify an image boundary line or boundary surface of all frames stored in the first frame memory 16; And an area discriminator 22 for discriminating an area by comparing image information of all previous frames stored in the second frame memory 18 and image information of all frames divided by the image divider 20 for each segment. An encoder with region discriminator, characterized in that it is added.