JPH09200772A - Compressed image data display device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To easily extract a desired image in a short time so that a moving image can be easily grasped. SOLUTION: This device is provided with a means (intra-extracter 2) for extracting an intra-frame compressed picture (I picture) from an inter-frame compressed video bit stream containing routine intra-frame compression like an MPEG system and a means for reducing the extracted I picture, etc., in the case of decoding such a video bit stream through a variable length decoder(VLD) 3, inverse quantizer 4 and inverter DCT equipment 5 or the like, this reduced image is written in frame memories (10 and 11) so that multi-display is enabled, and retrieval with the reduced picture as an index is facilitated.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a compact disc memory (CD-R) as a digital image medium.
From OM, for example, MPEG (Moving Pict)
The present invention relates to a compressed image data display device for decoding and displaying image data compressed by the ure Experts Group) method.

[0002]

2. Description of the Related Art FIG. 8 is a block diagram showing a general decoding system of this type, and FIG. 9 is an explanatory diagram of an MPEG system stream. As shown in FIG. 8, the processing device (CP
U) 20, CD-ROM decoder 21, MPEG video decoder 22, video D / A (digital / analog converter) 23, MPEG audio decoder 24, audio D / A 25 and the like. CD-R
The data output from the OM decoder 21 is the MPEG system stream 26 as shown in FIG. 9, which is a multiplexed packetized data of video and audio. A1, A2 ... Are audio packets, V1, V2 ...
Is a video packet, and padding is used to make the data length constant.

That is, the CPU 20 takes out the multiplexed system stream from the CD-ROM decoder 21, separates it into an audio stream and a video stream, and separates the separated video stream into an MPEG video decoder 22 and the audio stream into an MPEG audio decoder. Write to 24. The decoders 22 and 24 respectively perform decoding processing, and the video D / A 23,
Output to audio D / A 25.

FIG. 10 is a block diagram showing a general example of an MPEG video decoder. The input MPEG video stream has a VLD (variable length decoder) 3, an inverse quantizer 4 and an inverse DCT (IDCT: Inverse Di).
screen Cine Transform) 5
Each processing is performed at. That is, the VLD 3 decodes the variable-length coded data, converts the data into quantized data (quantized conversion coefficient) and outputs the quantized data, and outputs different signals depending on the picture type. Control the data path via 15. The inverse quantizer 4 inversely quantizes the quantized data input from the VLD 3 and outputs a DCT transform coefficient.
The inverse DCT unit 5 performs inverse DCT on the input transform coefficient and outputs data such as luminance Y, color differences Cr, Cb, etc., which become image data.

The control of the data path is different for each picture type, and the picture types are I picture and P.
There are pictures and B pictures. Figure 11 is a GOP
It is an explanatory view for explaining the composition of (group of pictures) and the prediction method of each picture. As shown in FIG. 11, the GOP includes I, P and B pictures, and
A picture (also referred to as an intra picture) is processed by intra-frame coding without referring to information of other images, and a P picture is predicted by using information of a past I picture or P picture, and is calculated as an inter-frame difference. Encoding process is performed. The B picture is predicted using the I and P pictures that are temporally preceding and following, and is encoded as an interframe difference. For the prediction of B pictures, the simple one is to simply average the preceding and following I and P pictures. Therefore, since the image to be referred to differs depending on each picture type, the data path becomes different.

FIG. 12 is an explanatory diagram of a data path for each picture type. FIG. 3A shows a data path of the I picture, and the output from the inverse DCT is stored in the prediction memory 1. As the output from the prediction memory 1, the I and P picture data written in the past are output at that timing. FIG. 2B shows a P picture data path, and the sum of the output from the prediction memory 1 and the output from the inverse DCT is written to the prediction memory 2. As the output from the prediction memory 2, the I and P picture data written in the past are output at that timing. In addition, FIG.
Writing to the prediction memories 1 and 2 in (b) is performed alternately. FIG. 3C shows a B picture data path, in which the data of I and P pictures from the prediction memories 1 and 2 are averaged, and the data is added to the output from the inverse DCT to obtain the video D / A section. Is output to. At that timing, the prediction memories 1 and 2 are not written.

FIG. 13 shows the data structure of an MPEG video stream. The uppermost layer is a sequence layer, which is a series of screen groups having the same attributes, which have the same screen size and image rate. The second layer is the GOP layer, which is the minimum unit of the screen that is the unit of random access. There is a picture layer as the third layer, which is an attribute common to one screen and corresponds to the coded picture type of the screen. The fourth layer is a slice layer and indicates information common to small screens obtained by dividing one screen into arbitrary lengths. There is a macroblock layer as the fifth layer, which shows information common to macroblocks obtained by dividing a slice. The sixth layer is a block layer, which corresponds to one DCT block (8 pixels × 8 lines).

In the MPEG video bit stream as described above, there is a high speed reproduction mode in addition to the normal reproduction mode. The concept is shown in FIG. This is a mode in which some GOPs in the sequence layer of the bitstream are skipped, only the I pictures in the skipped GOPs are taken out, and reproduced. With such a high-speed reproduction technique, the entire moving image can be grasped and reproduction from a desired image can be performed. In a video CD using the MPEG compression method, the menu screen is created in advance as a still image so that the entire moving image can be grasped. By selecting this menu screen, a desired image can be obtained.

[0009]

However, in the case of fast-forwarding to a desired screen by the high-speed playback method in the above-mentioned device, since the I pictures are distant from each other in time, so-called video equipment of the video device is displayed on the screen to be played back. There is no "movement" like in high-speed playback. Therefore, it is difficult to distinguish whether it is moving forward or backward in terms of time, and it is difficult to find a desired screen, and it takes time to search. Further, in the video CD, a menu screen is prepared so that the entire moving image can be grasped, but since this is created in advance, it is not possible for the user to create a new one. Therefore, an object of the present invention is to make it possible to search for a desired screen in a short time in order to easily grasp a moving image.

[0010]

In order to solve such a problem, according to the invention of claims 1 and 2, when decoding a video bit stream compressed between frames including regular intra-frame compression, this video is decoded. A means for extracting only the in-frame compressed screen from the bitstream, a means for reducing the extracted in-frame compressed screen, a means for writing a reduced image in the frame memory and multi-displaying are provided.
Write the compressed image in the frame to the frame memory,
According to the invention of claim 1, the operation is performed continuously in terms of time.
In the invention described in (1), it is performed only when a scene change is detected from the color distribution of the current image and the immediately preceding image.

Further, in the invention of claim 3, claims 1 and 2
5. In addition to the invention of claim 4, there is provided skip number setting means for setting the number of skipped intra-frame compressed screens.
In the invention, the in-frame compressed screen is taken out by using only the color distribution information of the limited part. Further, in addition to any one of the first to fourth aspects of the invention, an associating means for associating the reduced screen with the time code is added, and the decoding is started from the instructed time code position.

According to the invention of claim 1, a plurality of time-sequential reduced screens of intra images are displayed on the same screen. According to the invention of claim 2, the reduced screens of intra images are the same only when a scene change occurs. Multiple images are displayed on the screen. Further, according to the invention of claim 3, the time resolution between the reduced intra images can be freely changed, and according to the invention of claim 4, only when a change in the color distribution of a limited part between the intra images is detected. A plurality of reduced screens of intra images are displayed on the same screen, and according to the invention of claim 5, it is possible to perform a search using the reduced intra images as an index.

[0013]

1 is a block diagram showing a first embodiment of the present invention. In the figure, 1 is a time code extractor, 2 is an intra mode extractor, 3 is a VLD (variable length decoder), 4 is an inverse quantizer, 5 is an IDCT device, 6 is a comparator, 7 is a counter, 8 Is an adder, 9 is an address generator, 10 and 11 are prediction memories, 12 is an averager, and 13 and 1
Reference numerals 4 and 15 are switches (SW).

That is, the reduction ratio for resolution conversion,
Information such as ON / OFF of the intra mode indicating whether or not to select an intra image and a time code as an initial value for starting reduction reproduction is input from the outside. When the MPEG bit stream is input after the intra mode is set (ON) in the initial setting, the time code extractor 1
Thus, the time code in the currently input GOP (Group of Pictures) is output to the comparator 6. G
The OP has a start code (GSC) at the beginning, followed by a time code indicating the time from the start from the sequence. The time code output from the time code extractor 1 is compared with the time code initialized by the comparator 6, and if they match, a signal to that effect is sent to the address generator 9
Output to

The intra mode extractor 2 extracts only the I picture from the picture types, outputs it to the VLD 3, and outputs a count-up signal to the counter 7. In VLD3, inverse quantizer 4 and IDCT device 5, I
Only picture data is processed. The IDCT device 5 is initially set with a reduction ratio for resolution conversion.
The principle of resolution conversion is shown in FIG. Here, the reduction ratio is 1
An example of / 2, 1/4, 1/8 is shown. That is,
Reduction is realized by discarding high frequency data of DCT coefficients of 8 pixels × 8 lines according to the reduction ratio, performing inverse conversion of 4, 2, and 1 point, and then performing brightness adjustment. The IDCT device 5 outputs the data thus reduced.

The output from the IDCT device 5 is the prediction memory 1
Input to 0. Instead of the prediction memory, another frame memory may be used. The address generator 9 determines the offset position for multi-screen display based on the comparison match signal of the time code from the comparator 6, the signal from the counter 7 and the reduction ratio. After that, writing to the prediction memory 10 starts from each offset position according to the number of data from the IDCT device 5. FIG. 3 shows an example in which 16 I-pictures that are temporally continuous are displayed by reducing the size to 1/4.
The reduction processing by the IDCT device 5 is up to 1/8, but 1/16 and 1/32 can be realized by performing simple vertical and horizontal thinning after the reduction processing by the IDCT device.

FIG. 4 shows a second embodiment of the present invention.
The color distribution detector 1 is different from the first embodiment shown in FIG.
6 is added, and the count-up signal of the counter 7 is output from the color distribution detector 16. The color distribution detector 16 compares the histogram of the color distribution between the immediately previous reduced screen data and the currently input data to detect a difference in distribution. When the color distributions are significantly different, a count-up signal is output to the counter 7. FIG. 5 shows the principle of color distribution detection. Here, it is detected that a scene change has occurred at the position of the arrow. In other words, if there is not much change in what exists on the screen, the brightness histogram will be almost the same and will not be detected as a scene change. The difference in color distribution can be easily detected by obtaining a correlation value using a correlation calculation or the like and comparing this with a threshold value.

As described above, when there is no scene change, the offset address is overwritten without changing, and the offset address is updated only when the scene change occurs. Although the difference in color distribution is detected from one intra screen in the above, it may be detected by focusing only on a certain limited part of one sheet.

FIG. 6 shows a modification of FIG. As is apparent from the figure, this example is characterized in that a GOP skip unit 17 is added to the example shown in FIG. 4 and the number of skips can be set from the outside. That is, the GOP skip unit 17 skips according to the time code and the set value, and inputs only the skipped GOP to the intra mode extractor 2. As a result, it becomes possible to freely change the temporal resolution between the plurality of reduced-displayed intra images. That is, by making the time resolution rough, it is possible to observe the state of change over a long period of time. FIG. 7 shows a modification of FIG. This is characterized in that a GOP skip unit 17 is added to the one shown in FIG. Thereby, the reduced screen can be displayed at the determined time interval.

For each of the reduced screens that are multi-displayed as described above, a table or the like in which a time code can be drawn is prepared, and the time code can be output by pointing the reduced screen with a means such as a mouse. deep. Then, when the time code is output, if the normal MPEG reproduction is performed from the time code, the reduced screen created by the operator can be searched as an index. In the above description, the case of mainly decoding a video bit stream of the MPEG system has been described, but it goes without saying that the present invention can also be applied to the case of decoding a video bit stream similar to this.

[0021]

According to the present invention, the following effects can be expected. (1) According to the invention of claim 1, since it is possible to display the intra images which are temporally continuous, the relationship between the elapsed time and the screen can be grasped and the editing work and the like can be facilitated. (2) According to the invention of claim 2, the intra-image can be multi-displayed only when a scene change occurs, and each of the multi-displays has a cohesive scene of the moving image (the information on the screen changes). It will be one cut of a non-moving video) (3) According to the third aspect of the invention, the temporal resolution between the reduced intra images can be freely changed, and a long-time moving image can be grasped comprehensively. In addition, it is possible to grasp the details in detail after grasping the big picture. (4) According to the invention of claim 4, it is possible to detect a change in a certain limited portion between the intra images and perform multi-display, and it is possible to limit a place to be monitored in an application such as a monitoring image. The degree of change can be grasped as a reduced screen. As a result, it is possible to determine whether it is due to a change in sunshine or due to an intrusion of a person or the like. (5) According to the invention of claim 5, the moving image can be easily searched by only instructing the reduced screen.

[Brief description of drawings]

FIG. 1 is a configuration diagram showing a first embodiment according to the present invention.

FIG. 2 is an explanatory diagram of a resolution conversion method.

FIG. 3 is an explanatory diagram of intra extraction from a bitstream and multi-image display.

FIG. 4 is a configuration diagram showing a second embodiment according to the present invention.

FIG. 5 is an explanatory diagram of a color distribution detection method.

FIG. 6 is a configuration diagram showing a third embodiment according to the present invention.

FIG. 7 is a configuration diagram showing a fourth embodiment according to the present invention.

FIG. 8 is a block diagram showing an example of a general MPEG decoding system.

FIG. 9 is an explanatory diagram of an MPEG system stream.

FIG. 10 is a block diagram showing a general example of an MPEG video decoder.

FIG. 11 is an explanatory diagram illustrating a GOP configuration and a prediction method for each picture.

FIG. 12 is an explanatory diagram of a data path according to each picture type.

FIG. 13 is an explanatory diagram of a data structure of an MPEG video stream.

FIG. 14 is an explanatory diagram of a high speed reproduction mode.

[Explanation of symbols]

1 ... Time code extractor, 2 ... Intra mode extractor,
3 ... VLD (variable length decoder), 4 ... inverse quantizer, 5 ... I
DCT (Inverse Discrete Cosi)
ne Transform), 6 ... Comparator, 7 ... Counter, 8 ... Adder, 9 ... Address generator, 10, 11 ... Prediction memory, 12 ... Averager, 13, 14, 15 ... Switch (SW).

Claims (5)

[Claims] 1. A compressed image data display device for decoding and displaying an inter-frame compressed video bit stream including regular intra-frame compression, and an operation mode selection means for extracting only an intra-frame compressed screen, and Time code extracting means for extracting a time code from a video bit stream; decoding means for extracting and decoding an in-frame compressed screen from the video bit stream; reduction ratio setting means for setting a reduction ratio of the in-frame compressed screen; Counting means for counting the number of extracted in-frame compressed screens, comparing means for comparing the time code extracted by the time code extracting means with a preset time code, and the comparing means and the counting means. Based on each output and the set reduction ratio An offset address generating means for generating an offset address which becomes a display start position of the surface is provided, and when the operation mode is selected, an intra-frame compressed screen reduced from this offset address is written in a predetermined memory, and this can be multi-displayed. A compressed image data display device characterized by the above. 2. An operation mode selection means for extracting only an intra-frame compressed screen in a compressed image data display device for decoding and displaying a video bit stream compressed between frames including regular intra-frame compression, Time code extracting means for extracting a time code from a video bit stream; decoding means for extracting and decoding an in-frame compressed screen from the video bit stream; reduction ratio setting means for setting a reduction ratio of the in-frame compressed screen; A comparison means for comparing the time code extracted by the time code extraction means with a preset time code, and a scene change is detected from the color distribution of the current image and the immediately preceding image of the intra-frame compressed screen, and a count-up signal is output. Color distribution detection means and the count-up signal Counting operation means, and an offset address generating means for generating an offset address which is a display start position of the screen based on each output from the counting means and the comparing means and the set reduction ratio, and the operation mode A compressed image data display device characterized in that, when selected, an in-frame compressed screen reduced from this offset address is written in a predetermined memory so that it can be multi-displayed. 3. A skip number setting means for setting the number of skips of the intra-frame compressed screen to be skipped, and the time resolution of the multi-display intra-frame compressed screen can be arbitrarily changed according to the set number. The compressed image data display device according to claim 1 or 2. 4. The compressed image data display device according to claim 2, wherein the color distribution detecting means checks the color distribution of only a predetermined portion of the intra-frame compressed screen. 5. The compressed image data display according to claim 1, further comprising: associating means for associating the reduced screen with a time code, and decoding is started from a designated time code position. apparatus.