JP5346338B2 - Method for indexing video and apparatus for indexing video - Google Patents

Description

  The present invention relates to a method for indexing video and an apparatus for indexing video.

  Some picture processing applications use region of interest (ROI) detection to improve picture quality. For example, an encoding application often uses more resources to decode a region of interest and to encode the aforementioned region.

  Various methods allow detection of a region of interest in a picture. In particular, methods are known that take into account visual parameters and are based on the establishment of a saliency map of a picture or video that allows the definition of a region where the human eye will remain long when viewing the picture or video.

  Region-of-interest detection is currently primarily used to privilege the region of interest during encoding by providing more bandwidth (eg, by reducing the quantization step of the aforementioned region). Used before encoding.

  The appearance of mobile terminals (cell phones, PDAs, game consoles, portable DVD players, etc.), the development of display and screen techniques, and the rise of new services are all combined to display video on terminals with low display capacity. It is necessary. For example, since a television can be received on a mobile phone, there is a problem in displaying a high-density picture on a screen having a small size.

  The present invention primarily takes into account various applications rather than region of interest detection, and at least solves the problem of picture display on terminals with low display capacity, whether mobile terminals or not. It relates to the transmission of the aforementioned region of interest to a device or application.

For this purpose, the present invention proposes a method for indexing an encoded video data stream. According to the invention, the video data stream contains information about the location of the region of interest for each picture, and the method comprises:
Receiving an encoded video stream;
Encoding the video stream on the recording support;
Decoding position information of the region of interest;
Selecting a region of interest for each picture;
Decoding the video data;
Selecting a predetermined number of regions of interest in the video data stream from regions of interest selected for each picture;
Recording the selected region of interest.

In a preferred embodiment, during the recording process,
As selected and decrypted, the selected region of interest is recorded in temporary memory,
Once all the selected regions of interest have been recorded in temporary memory, the selected regions of interest are transferred to the permanent memory support (503).

  Preferably, before being recorded, the region of interest is formatted to obtain a uniform size for all selected regions of interest.

  Preferably, the method includes the step of encrypting the location of the region of interest with an encryption key.

  Preferably, the method includes the step of obtaining a decryption key by payment by the user.

  Preferably, the video data stream is encoded standard H.264. The position information is included in a supplementary extended information (SEI) type message.

  According to a preferred embodiment, SEI messages are encapsulated in real-time protocol packets (RTP) and RTP packets are encrypted.

  Preferably, the additional extended information type message regarding the region-of-interest position information is inserted into encoded data before and after each picture to be referred to.

According to a preferred embodiment, the location information is
The number of regions of interest in each picture,
Coordinates for each region of interest for each dimension of the picture,
A surface for each region of interest;
A weight for the importance of the region of interest relative to other regions of interest in the picture;
Information about content for each area of interest,
Information selected from any combination of the foregoing information is included.

  Preferably, the region of interest selection step for each picture selects the region of interest according to the weight for the importance of the region of interest.

  Preferably, the video coding standard uses a flexible macroblock arrangement, the region of interest is encoded into slices independently of other picture data, and the location information of the region of interest is encoded by the region of interest. The slice group number to be processed.

  Preferably, the additional extended information message includes an identifier for each slice group indicating whether or not it relates to one region of interest.

  Preferably, the method includes the further step of reading out the SEI message, and the step of decoding the video data decodes only the slice group including the region of interest.

The invention also relates to an apparatus for indexing an encoded video data stream. According to the invention, the video data stream contains information about the location of the region of interest for each picture,
Means for receiving an encoded video stream;
Means for recording the encoded video stream on a recording support;
Means for decoding position information of the region of interest;
Means for decoding the video data;
Means for selecting a region of interest for each picture;
Means for selecting a predetermined number of regions of interest in the video data stream from regions of interest selected for each picture;
Means for recording the selected region of interest.

  Detection of a region of interest in a picture is generally performed before encoding. This data is then used to facilitate encoding. The inventor of the present application recognizes that the position of the region of interest is of interest during decoding of a picture, particularly during display on a device with limited display capacity. In fact, the receiving terminal can choose to display only the region of interest and allow it to have better visibility of the aforementioned region for the complete picture display.

FIG. 2 is a diagram illustrating an encoding apparatus according to a preferred embodiment of the present invention. FIG. 3 is a diagram illustrating an encoding method according to a preferred embodiment of the present invention. FIG. 4 is a diagram illustrating a decoding apparatus according to a preferred embodiment of the present invention. FIG. 6 is a diagram illustrating a decoding method according to another embodiment of the present invention. FIG. 6 is a diagram illustrating a personal recording type apparatus according to another embodiment of the present invention. It is a figure which shows the indexing method in the apparatus of the personal recording type which implement | achieves the Example of this invention.

  The present invention will be more fully understood and illustrated by way of example and implementation with reference to the accompanying drawings, but not limiting enumeration.

  FIG. 1 illustrates the H.264 implementation of the preferred embodiment of the present invention. 1 shows an encoding device according to the H.264 / AVC encoding standard. In the preferred embodiment described above, the video stream is encoded.

The current frame F _n is supplied to the encoder input for encoding by the encoder. The aforementioned frames are encoded in the form of slices, i.e. the aforementioned frames are divided into subunits each containing a specific number of macroblocks corresponding to a 16x16 pixel group. Each macroblock is encoded in intra mode or inter mode. Whether in intra mode or inter mode, the macroblock is encoded based on the reconstructed frame. The module 109 determines the encoding mode of the current picture as the intra mode according to the content of the picture. In intra mode, P (shown in FIG. 2) is encoded, decoded, and reconstructed (uF′n in FIG. 2, u is unfiltered) of the current frame Fn. Includes sample. In inter mode, P results from motion estimation based on one or more F ′ _n−1 frames.

  The motion estimation module 101 establishes a motion estimation between the current frame Fn and at least one previous frame F'n-1. From this motion estimation, motion compensation module 102 generates frame P if the current picture Fn has to be encoded in inter mode.

  The subtracter 103 generates a signal Dn (that is, a difference between the picture Fn to be encoded and the picture P). This picture is then transformed in module 104 by DCT transformation. The transformed picture is then quantized by the quantization module 105. The picture is then reorganized by module 111. The CABAC (context based adaptive binary arithmetic coding) type entropy encoding module 112 then encodes each picture.

  Modules 106 and 107 (quantization module and inverse transform module, respectively) allow the difference D'n to be reconstructed after inverse quantization and inverse transformation following transformation and quantization.

  When the picture is encoded in intra mode by module 109, intra prediction module 108 encodes the picture. The uF'n picture and the sum of the D'n and P signals are obtained at the output 114 of the adder. This module 108 receives as input the reconstructed unfiltered F'n pictures.

  Filter module 110 may reconstruct from uF'n pictures to obtain filtered F'n pictures.

  The entropy decoding module 112 transmits a coded slice encapsulated with a NAL type unit. The NAL includes, for example, information about a header and a slice. The NAL type unit is transmitted to the module 113.

Module 116 allows a region of interest to be determined. Currently, several techniques allow a region of interest to be located in a picture. Techniques based on the establishment of a saliency map are particularly known. For example, a patent application filed on July 13, 2006 by Thompson Licensing, International Publication No. 2006/07263 (published July 13, 2006) is prominent. An effective method for creating a map is disclosed.

  Means 116 then establishes a saliency map for each picture of the video. In order to establish the aforementioned saliency map, parameters entered by the user may also be taken into account. For example, depending on the event that the video is related to, it is possible to define a certain important object of the scene that was filmed, especially in the case of a sporting event, that it is related to a soccer game It is. Effectively, this makes it possible to obtain a saliency map that weights saliency zones according to events. In soccer matches, it is preferable to focus on the ball, not the stand.

  The region of interest module thus allows one or more saliency zones (also represented as regions of interest) to be extracted. The aforementioned region of interest is then geographically located on the picture.

  The aforementioned region of interest is identified by coordinates according to the height and width of the picture. The aforementioned size can be extracted for each region of interest. It can also be associated with elements of semantic information. In fact, in the case of a soccer game, if the user is able to select a region of interest to display from a selection of several regions of interest to display, information about the region of interest may be required.

  Module 115 receives information regarding the region of interest for encoding into a SEI ("Additional Extended Information") type message.

  The SEI message is encoded as shown in the following table.

ｕｕｉｄ＿ｉｓｏ＿ｉｅｃ＿１１５７８：復号化器にメッセージ・タイプを示すための１２８ビットの単一のワード。

uuid_iso_iec_11578: A single 128-bit word to indicate the message type to the decoder.

user_data_payload_byte: 8 bits containing part of the SEI message
payloadSize = 17 (bytes), so the UUID is 16 and the unique data is 1.

  user_data_payload_byte:

ｎｕｍｂｅｒ＿ｏｆ＿ＲＯＩ：ピクチャ（又は後続ピクチャ）内に存在している関心領域の数
ｒｏｉ＿ｘ＿１６：１６個の画素の倍数での、関心領域のピクチャにおけるＸの位置
ｒｏｉ＿ｙ＿１６：１６個の画素の倍数で、関心領域のピクチャにおけるＹの位置
ｒｏｉ＿ｗ＿１６：関心領域のピクチャ内の幅（１６画素の倍数）
ｒｏｉ＿ｈ＿１６：関心領域のピクチャ内の高さ（１６画素の倍数）
ｓｅｍａｎｔｉｃ＿ｉｎｆｏｒｍａｔｉｏｎ：関心領域を特徴付ける題名
相対的重み：基本的に最も関心の高い関心領域が分かるようにピクチャの関心領域毎の重みを表す。

number_of_ROI: number of regions of interest present in the picture (or subsequent picture) roi_x_16: position of X in the region of interest in multiples of 16 pixels roi_y_16: multiple of 16 pixels in region of interest Y position in the picture roi_w — 16: width of the region of interest in the picture (multiple of 16 pixels)
roi_h — 16: height of the region of interest in the picture (multiple of 16 pixels)
semantic_information: Title that characterizes the region of interest Relative weight: Basically represents the weight of each region of interest in the picture so that the region of interest of highest interest is known.

  Macroblock_alignment: Represents the size of the region of interest in the number of starting macroblocks, the number of macroblocks, the width, and the height where the region of interest exists.

When a region of interest is detected using a saliency map, a saliency rate is obtained for each region of interest, and if the saliency is higher than a specific threshold previously determined by the technique for obtaining the saliency map Regions are classified as prominent. Therefore, in the SEI message, the region of interest is classified in ascending order of the saliency of all the regions that are more saliency than the fixed threshold.
Module 113 inserts the SEI message into the data stream and sends the encoded video stream to the transmission network.

  The SEI message is transmitted before each referenced picture.

  In other embodiments, the SEI message can be transmitted only when the position of at least one region of interest varies between two or more pictures. Thus, during decoding, the decoder will be in the immediate vicinity whether it is immediately before the picture to be decoded or if it is related to a previously received picture if the current picture does not precede the SEI message. Take into account the received SEI message.

  FIG. 2 illustrates the H.264 implementation of the preferred embodiment of the present invention. 2 illustrates an encoding method according to the H.264 / AVC encoding standard.

  During step E1, a saliency map associated with the video to be broadcast is determined. In order to determine the aforementioned saliency map indicating the region of interest, information about the video content may also be received to take this information into account during establishment of the saliency map. In particular, during a sporting event, the position of the ball corresponds to the user's region of interest, in which case the picture zone in which the ball is located is privileged. If the video corresponds to the broadcast of a television broadcast report, the region of interest may be determined, for example, by privileged the zone containing the moderator by detecting faces using known picture processing techniques. .

  At the end of the E1 process, one or more regions of interest for the video content are thus obtained.

  During step E2, the coordinates of the region of interest in the picture are determined. The size of the region of interest can also be obtained in pixels, and the semantic information on the content can be associated with each region of interest.

  In parallel, during step E3, the video stream is H.264. It is encoded according to the H.264 encoding standard. During encoding, zones detected as regions of interest are privileged. In order to privilege the region of interest at the coding level, a lower quantization step is applied.

  Subsequent to step E2, during step E4, an SEI message is created from the semantic information and location associated with the region of interest. The SEI message thus created is based on the SEI message described above in Tables 1 and 2.

  During step E5, the stream is H.264. In order to obtain an encoded stream according to the H.264 standard, it is constructed by inserting an SEI message into the stream.

  The encoded video stream is then transmitted to the decoding device in real time or delayed during step E6, which may be local or remote.

  FIG. The H.264 / AVC coding standard represents a preferred embodiment of the decoding device according to the invention.

  Module 209 receives the SEI message at the input. Module 209 extracts separate SEI messages. The useful data NAL is transmitted to the entropy decoding module 201.

  The SEI message is parsed by module 210. This module enables the decoding of the content of the SEI message representing the region of interest. The region of interest of each picture is thus identified in a simple manner and at the level of the decoding device before decoding each picture using information contained in the field macroblock_alignment.

  The macroblock is sent to the rearrangement module 202 to obtain a set of coefficients. The aforementioned coefficients undergo an inverse quantization in module 203 and an inverse DCT transform in module 204. At the output of module 204, D'n macroblocks are obtained, where D'n is a modified version of Dn. The predicted block P is added to D′ n by the adder 205 to reconstruct the macroblock uF′n. The block P is a block of the macroblock uF′n by the module 207 during the inter mode encoding, after the motion compensation performed by the module 208 of the preceding decoded frame, or in the case of intra mode encoding. Obtained after intra prediction. Filter 206 is applied to signal uF'n to reduce the effects of distortion, and a reconstructed frame F'n is generated from the series of macroblocks.

  Using information about the region of interest included in the SEI message, blocks representing the region of interest are detected in the stream before display, the aforementioned blocks are identified, cropped according to the user's choice, such as a PDA or mobile phone Can be sent to other devices for display.

  For example, by inputting semantic information, it is possible to leave the selection to the user to select a macroblock to be displayed. For example, “Ball” is input, and in this case, a region of interest including the ball is displayed. If there is no region of interest associated with this semantic information, it is possible to display the entire region of interest. It is possible to display various regions of interest on the screen in the form of a mosaic. When a single region of interest is displayed, the region of interest is displayed on the screen in a zoom so as to occupy the entire screen.

  The decoding device thus only decodes macroblocks that are likely to contain interest information for the user. In this way, the decoding is faster and requires less resources at the level of the decoding device and therefore on the receiving side. This is particularly effective when the receiving device is a mobile terminal with limited processing capacity.

  FIG. 4 shows an H.264 implementation of the preferred embodiment of the present invention. 2 shows a decoding method according to the H.264 / AVC coding standard.

  The above-described method can be realized in a mobile terminal having a limited display capacity.

  During step S1, the required display type is selected. The selection is made by means of a user interface that exists on the mobile terminal. If it is decided to work in full picture mode, the entire video stream is displayed as it is transmitted by the transmitter. Alternatively, only the region of interest of the picture is displayed. This particular mode constitutes a feature of the present invention. If the region of interest is to be displayed, the process moves to step S2, otherwise the process moves to step S8. Various SEI messages can be inserted into the video streams of other applications, in which case there may be a step of SEI message analysis before or during step S8.

During step S2, the user selects how to use the region of interest. In particular, users
The maximum number of regions of interest you want to display,
A mode of displaying various regions of interest on the screen (for example, mosaic format),
It is possible to select the degree of zoom desired for the region of interest,
Using keywords, it is possible to select a region of interest whose "semantic information" field contains the keyword. In this case, for each picture, it is necessary to display a single region of interest (in this case, the region of greatest saliency) for each picture containing the keyword, or several regions of interest containing the keyword It is also possible to define whether it is necessary to display.

  During step S3, SEI messages present in the stream are analyzed as they are received. The SEI message is used to encode the position of the region of interest of the picture detected before picture encoding. Thus, for each picture, there may be one or more regions of interest, depending on the visual characteristics of the picture, or by the picture content, or both. The SEI message is encoded according to Tables 1 and 2 described above. Information about the SEI message is recorded until the display of the corresponding picture in time.

  During step S4, all pictures are decoded according to the decoding standard.

  During step S5, the decoded region of interest is processed according to what the user selected during step S2. If the user chooses to zoom the main region of interest of the picture, the zone is enlarged to reach the maximum display size during step S6. If the user selects a mosaic of regions of interest, the picture is reconstructed with regions of interest, and each region of interest is magnified by the number of regions of interest selected for display and the screen size. When the user defines a keyword, the region of interest containing the keyword is displayed and zoomed.

  During step S7, the region of interest is displayed on the screen of the mobile terminal according to the user's desire.

  During step S8, following the deselection by the user to display only the region of interest, the entire video stream is decoded for display.

  FIG. 5 illustrates the video indexing application of the present invention.

  FIG. 5 partially shows a personal recorder (PVR) type device 500. PVR 500 receives a compressed video stream at its input. According to the embodiment described above, this video data stream is H.264. H.264 encoding standard. The compressed video stream specifically includes SEI messages as described above in Tables 1 and 2.

  This video data stream is partially transmitted to the recording support 503. The recording support can be considered a hard disk, holographic support, memory card, or “blue ray” disk. This recording support may be remote in other embodiments.

  The video data stream is transmitted in another part to decoder 501 for decoding in real time (eg, for display on a television receiver). In known devices, the stream is sent to the decoder 501 when the user wants to see it in real time. In the negative case, if recording is requested, it is not decrypted but simply recorded.

  The present invention, according to this aspect, provides for decoding a portion of a video data stream even when real-time viewing is not required. In the case of parts of a video stream, in particular a region of interest or a specific region of interest.

  When the decoder 501 receives the video stream requested to be recorded, the data is sent to the recording support 503. The recording support 503 records data as it is received. At the same time, decoder 501 receives the video data stream and progressively decodes the SEI message. The decoded region of interest is sent to the video indexing module 502 which serves to temporarily record before sending it to the recording support 503.

  FIG. 6 shows the method implemented by the decoder 501 and the indexing module 502.

  During step T1, the video data stream is received by the decoder 501. During step T2, the decoder 501 decodes the SEI message present in the video data stream. The decrypted SEI message is an SEI message as described above in Tables 1 and 2. The decoder can also decode other SEI messages, but this is not the object of the present invention. Each SEI message may represent one or more regions of interest for each picture as described above in Tables 1 and 2. During step T3, the decoder 501 analyzes each SEI message and decodes each picture. During this process, the weight indicated in the SEI message is used to select which region of interest is recorded for each picture. In the preferred embodiment, the region of interest with the highest saliency (ie, the highest weight) is maintained.

  Once the region of interest is decoded, it is sent to the indexing module 502 during step T4. Interest in recording the region of interest for each picture (all pictures) is low. This is because the amount of information is large and efficient video indexing is not possible. Thus, the indexing module determines which picture to use to index the video. According to the preferred embodiment described above, only about 10 pictures are selected for a 1.5 hour video. In other embodiments, the number of pictures is assumed to be large. The above ten pictures are taken at regular intervals. The selected picture is temporarily recorded in a RAM type memory (not shown) included in the indexing module 502. In order to be displayed in the best way, the picture is zoomed during step T5. That is, they are enlarged so that they are all the same size. In the preferred embodiment, this size may be the size of the picture. To that end, they are read in a temporary memory and re-recorded after expansion. In another embodiment, the picture is enlarged before recording in temporary memory.

  According to another embodiment, the image is presented as a mosaic on the display. Thus, instead of being enlarged, the image is reduced to one single size for all as well.

  When the entire video is received and recorded at the recording support 503, the indexed picture is transferred from the temporary memory to the recording support 503 and recorded in a file.

Then, according to the desired use, the region of interest is used for indexing . By the indexing, if the user wants to query the contents of the database, the recorder it is capable ing for displaying an image of the video.

  According to another aspect of the present invention, it is possible to encrypt the position data of the region of interest during the encoding of the SEI message. Thus, only the user with the decryption key can access the region of interest, access the visualization of the region of interest, or access the index of the video stream with the location information of the region of interest. This encryption step (see FIG. 2) becomes step E4 '(not shown), but is inserted after step E4.

  The acquisition of the decryption key can be, for example, a pay service target from a program broadcaster.

  To do this, the SEI message for the region of interest is encapsulated in RTP (Real Time Protocol) type packets and sent on another video port. The temporal CTS type label can relate the corresponding picture to the SEI message for the region of interest. Effectively, this transmission mode makes it possible to encrypt only RTP packets that contain SEI messages, not video.

  Decoding is performed at the terminal receiver level.

  In the case of MPEG-2 TS encapsulation, the encryption standard used is DVB-CSA, and the SEI message for the region of interest is encapsulated in a PID different from that of the video. The SEI message for the region of interest is related to the corresponding picture via the PTS (Time Stamp) in the PES packet header. This transmission mode allows encryption of only PIDs containing SEI messages regarding the region of interest, not video PIDs.

  In another embodiment, the video stream is encoded in H.264 using FMO (Flexible Macroblock Array), which allows independent portions of pictures to be encoded independently and thus decoded independently. It is encoded according to the H.264 / AVC encoding standard. The FMO mode uses “slice group”. The “slice group” is defined in the standard. In this embodiment, the region of interest is encoded in a different group than the rest of the picture. The PPS type NAL includes a map of “slice groups”. An SEI message such as that described below is inserted, indicating a “slice group” in which the region of interest is encoded.

  The following table shows the format of the SEI message used by this example.

ｕｕｉｄ＿ｉｓｏ＿ｉｅｃ＿１１５７８：復号化器にメッセージ・タイプを示すための１２８ビットの単一のワード。

uuid_iso_iec_11578: A single 128-bit word to indicate the message type to the decoder.

user_data_payload_byte: 8 bits containing part of the SEI message
payloadSize = 17 (bytes). Therefore, the UUID is 16 and the unique data is 1.

  user_data_payload_byte:

Ｓｌｉｃｅ＿ｇｒｏｕｐ（ｉ）＿ｉｄ： ｓｌｉｃｅ＿ｇｒｏｕｐ＿ｉｄが「１」の場合、ｓｌｉｃｅ＿ｇｒｏｕｐは関心領域を表し、ｓｌｉｃｅ＿ｇｒｏｕｐ＿ｉｄが「０」の場合、ｓｌｉｃｅ＿ｇｒｏｕｐはピクチャの残りを表す。

Slice_group (i) _id: When slice_group_id is “1”, slice_group represents a region of interest, and when slice_group_id is “0”, slice_group represents the rest of the picture.

  For each slice_group representing a region of interest, semantic information, relative weights, and related macroblocks can be defined.

  Thus, only the macroblock corresponding to the region of interest can be decoded during reception as it is independently identified and encoded.

Claims (14)

A method for indexing an encoded video data stream by a personal recorder device , wherein the video data stream includes information about the location of a region of interest in each picture, the method comprising:
The personal recorder device receiving an encoded video data stream for a decoder;
Recording support recording the encoded video data stream on the recording support;
A decoder decoding position information of the region of interest;
A step of video indexing module selects the region of interest for each picture,
A step of the decoder to decode the encoded video data stream,
The step of the video indexing module selects a predetermined number of regions of interest of the video data stream from said selected region of interest for each picture,
The recording support comprising: recording the selected region of interest of the video data stream on the recording support. The indexing method of claim 1, wherein the recording support records the selected region of interest of the video data stream.
As the video data stream is decoded and the region of interest of the video data stream is selected, the selected region of interest of the video data stream is recorded in temporary memory;
The indexing method, wherein when the selected region of interest of the video data stream is recorded in the temporary memory, the selected region of interest of the video data stream is transferred to a recording support. The indexing method according to claim 1, wherein the method comprises:
Between the video indexing module selecting a predetermined number of regions of interest in the video data stream and the recording support recording the selected regions of interest in the video data stream. Indexing comprising zooming the selected region of interest of the video data stream with the personal recorder device to obtain a uniform size of the selected region of interest of the video data stream how to. 4. The indexing method according to claim 1, wherein the personal recorder device encrypts the position information of the region of interest with an encryption key. . 5. The indexing method according to claim 4, comprising the step of obtaining a decryption key by the personal recorder device by payment by a user.   6. A method for indexing as claimed in any preceding claim, wherein the video data stream is H.264. A method of indexing, encoded according to the H.264 / AVC encoding standard, wherein the location information is included in a supplemental extension information (SEI) type message.   The indexing method according to claim 5 or 6, wherein the supplemental extension information (SEI) type message is encapsulated in a real-time protocol packet (RTP) and the RTP packet is encrypted. how to.   The indexing method according to any one of claims 5 to 6, wherein an additional extension information (SEI) type message related to region-of-interest position information is provided for each picture referred to by the additional extension information type message. A method of indexing that is inserted into a pre- or post-encoded video data stream. The indexing method according to claim 1, wherein the position information is
The number of regions of interest in each picture,
The coordinates of each region of interest according to picture width and picture height,
The width and height of each region of interest;
The relative weights of the associated cardiac region to other areas of the picture,
A method of indexing comprising information about the content of each region of interest and information selected from any combination of information. A method of indexing according to any one of claims 1 to 9, the step of the video indexing module selects a region of interest for each of the picture, the video indexing module region of interest A method of indexing, wherein regions of interest are selected according to the relative weights of the regions of interest. A method for indexing according to any one of claims 6 to 10, comprising:
The video coding standard uses a flexible macroblock arrangement, the region of interest is encoded into slice groups independently of other picture data, and the location information of the region of interest is encoded in the region of interest. Indexing method including the slice group number.   12. The indexing method of claim 11, wherein the supplemental enhancement information (SEI) type message includes an identifier that indicates for each slice group if it is related to a region of interest. 13. The indexing method of claim 12, further comprising the step of the personal recorder device reading the supplemental enhancement information (SEI) type message, wherein the decoder is the encoded video data stream. Is a method of indexing, in which the decoder decodes only the slice group indicated by the identifier that the slice group relates to one region of interest. An apparatus for indexing an encoded video data stream, wherein the video data stream includes information regarding the location of a region of interest for each picture, the apparatus comprising:
Means for receiving an encoded video data stream;
Means for recording the encoded video data stream on a recording support;
Means for decoding position information of the region of interest;
Means for decoding the video data stream;
Means for selecting a region of interest for each picture;
Means for selecting a predetermined number of regions of interest in the video data stream from regions of interest selected for each picture;
Means for recording the selected region of interest of the video data stream.

Images