Video transmission method and terminal

Abstract

The application discloses a video transmission method and a terminal, the method comprises the following steps: acquiring a source video stream, and acquiring a key frame from the source video stream; determining a key data packet and other data packets in the key frame based on a preset analysis inference model; respectively encoding and processing the key data packet and the other data packets based on different resolutions to obtain a first data packet and a second data packet, wherein the resolution of the second data packet is lower than that of the first data packet; transmitting the first data packet through a reliable transmission protocol to obtain a first data packet set; transmitting the second data packet through an unreliable transmission protocol to obtain a second data packet set; and splicing the key frame according to the first data packet set and the second data packet set to obtain a target video for playing. The application can guarantee the quality of a video picture while reducing the data transmission amount and effectively improving the data transmission efficiency.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of video processing, and in particular to a video transmission method and terminal. BACKGROUND

[0002] At present, most video streams and various live video streams provide multiple resolution functions. Under different resolutions, the server will use different data transmission functions. For example, the commonly used resolutions at present include 720P, 1080P, 2K, 4K and 8K. These resolutions are all progressively increased according to video clarity, that is, the higher the resolution, the clearer the video, and the higher the resolution, the more data the video transmission requires, which causes the video transmission to be time-consuming, the network IO (input / output) bandwidth occupied by the server to be large, and the user to need a certain waiting time to play the video, resulting in poor video playing experience. Therefore, under normal circumstances, the user decides to select which resolution to use for data transmission according to the current network transmission state and viewing demand. SUMMARY

[0003] The technical problem to be solved by the present application is to provide a video transmission method and terminal, which can reduce the video data transmission amount and improve the video transmission efficiency.

[0004] To solve the above technical problems, the technical scheme adopted by the present application is:

[0005] A video transmission method, comprising:

[0006] obtaining a source video stream and obtaining a key frame from the source video stream;

[0007] determining key data packets and other data packets in the key frame based on a preset analysis and inference model;

[0008] encoding and processing the key data packets and the other data packets based on different resolutions to obtain first data packets and second data packets, the resolution of the second data packets being lower than the resolution of the first data packets;

[0009] transmitting the first data packets through a reliable transmission protocol to obtain a first data packet set;

[0010] transmitting the second data packets through an unreliable transmission protocol to obtain a second data packet set;

[0011] splicing the key frame based on the first data packet set and the second data packet set to obtain a target video for playing.

[0012] To solve the above technical problems, another technical scheme adopted by the present application is:

[0013] A video transmission terminal comprises a memory, a processor and a computer program stored on the memory and running on the processor, and the processor implements the following steps when executing the computer program:

[0014] Obtaining a source video stream and a key frame from the source video stream;

[0015] Determining a key data packet and other data packets in the key frame based on a preset analysis inference model;

[0016] Encoding and processing the key data packet and the other data packets based on different resolutions respectively to obtain a first data packet and a second data packet, the resolution of the second data packet being lower than that of the first data packet;

[0017] Transmitting the first data packet through a reliable transmission protocol to obtain a first data packet set;

[0018] Transmitting the second data packet through an unreliable transmission protocol to obtain a second data packet set;

[0019] Splicing the key frame based on the first data packet set and the second data packet set to obtain a target video for playing.

[0020] The application has the beneficial effects that the key frame is divided into the key data packet and the other data packets based on the analysis inference model to distinguish the key data and non-key data in the video. Meanwhile, the key data packet and the other data packets are encoded and processed respectively to obtain the first data packet and the second data packet with different resolutions, based on which, the first data packet with higher resolution is used for the key part of the video to ensure the definition, and the second data packet with lower resolution is used for the non-key part of the video to reduce the data amount. In addition, the reliable transmission protocol is used for transmitting the first data packet to ensure the data quality, and the unreliable transmission protocol is used for transmitting the second data packet, so that even if the second data packet is lost, it does not need to be retransmitted. In this way, the data transmission amount is reduced while the video picture quality is ensured, and the data transmission efficiency is effectively improved. BRIEF DESCRIPTION OF DRAWINGS

[0021] Figure 1 A step flowchart of a video transmission method provided for the embodiment of the application;

[0022] Figure 2 A structural schematic diagram of a video transmission terminal provided for the embodiment of the application;

[0023] LABEL EXPLANATION

[0024] 100, a video transmission terminal; 101, a memory; 102, a processor. DETAILED DESCRIPTION

[0025] To make the technical contents of the present application, the purposes achieved and the effects more comprehensible, the following will describe the embodiments in conjunction with the accompanying drawings.

[0026] The embodiment of the present application provides a video transmission method, comprising:

[0027] Obtaining a source video stream, and obtaining a key frame from the source video stream;

[0028] Determining key data packets and other data packets in the key frame based on a preset analysis and inference model;

[0029] Encoding and processing the key data packets and the other data packets based on different resolutions respectively to obtain first data packets and second data packets, the resolution of the second data packets being lower than that of the first data packets;

[0030] Transmitting the first data packets through a reliable transmission protocol to obtain a first data packet set;

[0031] Transmitting the second data packets through an unreliable transmission protocol to obtain a second data packet set;

[0032] Splicing the key frame based on the first data packet set and the second data packet set to obtain a target video for playing.

[0033] From the above description, the beneficial effects of the present application are that the key frame is divided into key data packets and other data packets based on an analysis and inference model to distinguish key data and non-key data in the video. Meanwhile, the key data packets and the other data packets are encoded and processed respectively to obtain first data packets and second data packets of different resolutions. Based on this, the first data packets of higher resolution are used for the key part of the video to ensure the definition, and the second data packets of lower resolution are used for the non-key part of the video to reduce the data amount. In addition, the reliable transmission protocol is used to ensure the data quality when transmitting the first data packets, and the unreliable transmission protocol is used when transmitting the second data packets, so that even if the second data packets are lost, retransmission is not needed. In this way, the data transmission amount is reduced while the video picture quality is ensured, and the data transmission efficiency is effectively improved.

[0034] Further, the determination of the key data packets and the other data packets in the key frame based on the preset analysis and inference model comprises:

[0035] Identifying key pictures in the key frame based on the preset analysis and inference model to obtain key picture blocks and other picture blocks;

[0036] Marking frame data corresponding to the key picture blocks in the key frame as key data packets;

[0037] frame data corresponding to the other picture blocks in the key frame is marked as other data packets.

[0038] As can be seen from the above description, the current video playing needs to load all video data before playing, causing a delay in playing, and the user experience is poor. However, since the user is not interested in all the content in the entire video, the key frame is divided into a user-focused key picture block and other picture blocks, so that the frame data corresponding to the picture blocks is encoded at different resolutions, which not only ensures the picture quality, but also effectively reduces the overall data volume of the video, and different data packets are transmitted using different transmission protocols, further improving data transmission efficiency and reducing data delay.

[0039] Further, the analysis and inference model is trained based on the user's attention to different picture positions in the key frame;

[0040] The identification of the key picture in the key frame based on the preset analysis and inference model to obtain the key picture block and the other picture block includes:

[0041] Obtaining a display picture of the key frame, and dividing the display picture into a plurality of picture blocks according to picture positions;

[0042] Based on the preset analysis and inference model, the target of the plurality of picture blocks in the key frame is identified to obtain the key picture block and the other picture block of the key frame.

[0043] As can be seen from the above description, the analysis and inference model is trained based on the user's attention to different picture positions in the key frame, so that the analysis and inference model can effectively identify the key picture in the key frame that the user pays more attention to, and different video transmission methods are used for the key picture and other pictures, ensuring the video viewing experience while improving the data transmission efficiency.

[0044] Further, the encoding processing of the key data packet and the other data packet based on different resolutions to obtain the first data packet and the second data packet includes:

[0045] Receiving a first resolution selected by the user, and encoding the key data packet based on the first resolution to obtain a first data packet;

[0046] Obtaining a second resolution lower than the first resolution, and encoding the other data packet based on the second resolution to obtain a second data packet.

[0047] As can be known from the above description, in the case that the user has selected a fixed resolution, the fixed resolution is used for data transmission of the key data packet which is focused by the user, and a lower resolution is used for data transmission of other data packets which are not particularly focused by the user, compared with the method that all data is transmitted by using a fixed resolution, the method of using different resolutions for transmission can not only ensure the picture instruction, but also effectively reduce the data transmission amount.

[0048] Further, the transmitting the first data packet by using the reliable transmission protocol further comprises:

[0049] marking frame data in the first data packet which is adjacent to the other picture block and located in a preset range as edge frame data;

[0050] marking frame data in the first data packet which is not adjacent to the other picture block or located out of the preset range as center frame data;

[0051] the transmitting the first data packet by using the reliable transmission protocol comprises:

[0052] transmitting the center frame data by using the reliable transmission protocol and transmitting the edge frame data by using the unreliable transmission protocol to obtain a first data packet set.

[0053] As can be known from the above description, since the first data packet and the second data packet are transmitted by using different resolutions respectively, when the first data packet and the second data packet are spliced, there may be a visual boundary line in the video picture. In order to avoid the video picture from having an obvious visual boundary line, the center frame data in the first data packet is transmitted by using the reliable transmission protocol, and the edge frame data in the first data packet is transmitted by using the unreliable protocol, so that the center frame data can be completely transmitted to ensure the picture resolution, and the edge frame data adjacent to the second data packet is allowed to be lost and does not need to be transmitted to reduce the data transmission amount, so that the first data packet and the second data packet are spliced more smoothly, and the quality of the video picture is optimized.

[0054] Another embodiment of the present application provides a video transmission terminal, comprising a memory, a processor and a computer program stored in the memory and running on the processor, and the processor implements the following steps when executing the computer program:

[0055] obtaining a source video stream and obtaining a key frame from the source video stream;

[0056] determining a key data packet and other data packets in the key frame based on a preset analysis and inference model;

[0057] Encode and process the key data packet and the other data packet based on different resolutions respectively to obtain a first data packet and a second data packet, wherein the resolution of the second data packet is lower than that of the first data packet;

[0058] Transmit the first data packet through a reliable transmission protocol to obtain a first data packet set;

[0059] Transmit the second data packet through an unreliable transmission protocol to obtain a second data packet set;

[0060] Play the target video obtained by splicing the key frame based on the first data packet set and the second data packet set.

[0061] From the above description, the beneficial effects of the present application are that the key frame is divided into a key data packet and an other data packet based on an analysis and inference model to distinguish between key data and non-key data in the video. Meanwhile, the key data packet and the other data packet are encoded and processed to obtain a first data packet and a second data packet with different resolutions. Based on this, the key part of the video adopts the first data packet with a higher resolution to ensure the definition, while the non-key part of the video adopts the second data packet with a lower resolution to reduce the data volume. In addition, the reliable transmission protocol is adopted when transmitting the first data packet to ensure the data quality, and the unreliable transmission protocol is adopted when transmitting the second data packet, so that even if the second data packet is lost, it does not need to be retransmitted. In this way, while ensuring the quality of the video picture, the data transmission volume is reduced, and the data transmission efficiency is effectively improved.

[0062] Further, the determination of the key data packet and the other data packet in the key frame based on the preset analysis and inference model comprises:

[0063] Identifying the key picture in the key frame based on the preset analysis and inference model to obtain a key picture block and an other picture block;

[0064] Marking the frame data corresponding to the key picture block in the key frame as a key data packet;

[0065] Marking the frame data corresponding to the other picture block in the key frame as an other data packet.

[0066] As can be known from the above description, the current video playing needs to load all video data before playing, causing a playing delay and a poor user experience. However, since the user does not pay attention to all the content in the entire video, the key frame is divided into a key picture block and other picture blocks that the user pays attention to, so that the frame data corresponding to the picture blocks is encoded at different resolutions, which can not only ensure the picture quality, but also effectively reduce the overall data volume of the video, and different data packets are transmitted by using different transmission protocols, further improving the data transmission efficiency and reducing the data delay.

[0067] Further, the analysis inference model is trained based on the attention degree of the user to different picture positions in the key frame;

[0068] The key picture in the key frame is identified based on the preset analysis inference model to obtain a key picture block and other picture blocks, including:

[0069] An exhibition picture of the key frame is obtained, and the exhibition picture is divided into a plurality of picture blocks according to picture positions;

[0070] The plurality of picture blocks in the key frame are target-identified based on a preset analysis inference model to obtain a key picture block and other picture blocks of the key frame.

[0071] As can be known from the above description, the analysis inference model is trained based on the attention degree of the user to different picture positions in the key frame, so that the analysis inference model can effectively identify the key picture in the key frame that the user pays more attention to, and different video transmission methods are used for the key picture and other pictures, thereby ensuring the video watching experience and improving the data transmission efficiency.

[0072] Further, the key data packet and the other data packet are encoded and processed based on different resolutions to obtain a first data packet and a second data packet, including:

[0073] A first resolution selected by the user is received, and the key data packet is encoded and processed based on the first resolution to obtain a first data packet;

[0074] A second resolution lower than the first resolution in resolution level is obtained, and the other data packet is encoded and processed based on the second resolution to obtain a second data packet.

[0075] As described above, when the user has selected a fixed resolution, the user focuses on transmitting data packets at a fixed resolution, while other data packets that the user is not particularly concerned about are transmitted at a lower resolution. Compared to transmitting all data at a fixed resolution, this method of transmitting at different resolutions not only ensures the display instructions but also effectively reduces the amount of data transmitted.

[0076] Furthermore, the step of transmitting the first data packet via a reliable transport protocol further includes:

[0077] Frame data in the first data packet that is adjacent to the other picture blocks and located within a preset range is marked as edge frame data;

[0078] Frame data in the first data packet that is not adjacent to other screen blocks or is located outside a preset range is marked as center frame data;

[0079] The step of transmitting the first data packet through a reliable transmission protocol to obtain the first data packet set includes:

[0080] The center frame data is transmitted using the reliable transmission protocol, and the edge frame data is transmitted using an unreliable transmission protocol to obtain the first data packet set.

[0081] As described above, since the first and second data packets are transmitted at different resolutions, a visual boundary line may appear in the video frame when the first and second data packets are stitched together. To avoid obvious visual boundary lines in the video frame, the center frame data in the first data packet is transmitted using a reliable transmission protocol, while the edge frame data is transmitted using an unreliable protocol. This ensures that the center frame data is transmitted completely to maintain the image resolution, while edge frame data adjacent to the second data packet is allowed to be dropped without transmission to reduce the amount of data transmitted. This makes the stitching of the first and second data packets smoother and optimizes the quality of the video frame.

[0082] This invention provides a video transmission method and terminal, applicable to video transmission between a server and a client. It reduces the amount of video data transmitted, thereby improving video transmission efficiency. Specific embodiments are described below:

[0083] Please refer to Figure 1 Embodiment 1 of the present invention is as follows:

[0084] A video transmission method, comprising:

[0085] S1. Obtain the source video stream and extract keyframes from the source video stream.

[0086] In some embodiments, a set of consecutive image frames from the source video stream are obtained as keyframes.

[0087] S2. Based on a preset analysis and reasoning model, determine the key data packets and other data packets in the key frame.

[0088] Specifically, step S2 includes:

[0089] S21. Based on a preset analysis and reasoning model, identify key images in the key frames to obtain key image blocks and other image blocks.

[0090] In an optional implementation, the analysis and inference model is trained based on the user's attention to different screen positions in the keyframe, then step S21 includes:

[0091] S211. Obtain the display screen of the key frame and divide the display screen into multiple screen blocks according to the screen position.

[0092] In some embodiments, the display screen of each keyframe is divided into 5 screen blocks according to the 5 screen positions of the middle, top left, bottom left, top right and bottom right.

[0093] S212. Based on a preset analysis and reasoning model, target recognition is performed on multiple screen blocks in the key frame to obtain the key screen blocks and other screen blocks of the key frame.

[0094] In some embodiments, the number of key frame blocks can be selected based on the screen area occupied by the key frame. For example, if the current analysis and inference model detects that the screen area occupied by the key frame includes a middle frame block and a lower left frame block, then the middle frame block and the lower left frame block can be set as key frame blocks of the key frame, while the upper left frame block, the upper right frame block, and the lower right frame block are automatically marked as other frame blocks of the key frame. If the current analysis and inference model detects that the screen area occupied by the key frame only includes a middle frame block, then the middle frame block is directly set as the key frame block of the key frame, while the upper left frame block, the lower left frame block, the upper right frame block, and the lower right frame block are automatically marked as other frame blocks of the key frame.

[0095] In some embodiments, each keyframe is analyzed and identified based on a preset analysis and reasoning model, which of the five image blocks is the key image of the current keyframe. Since a keyframe is a set of consecutive image frames, and the key image in a set of consecutive adjacent keyframes does not have a significant positional change, the key image blocks in a set of consecutive adjacent keyframes are usually the same. For example, the key image blocks of keyframes 0-10 in the source video stream are all the lower left image blocks, while the key image blocks of keyframes 11-20 in the source video stream are all the middle image blocks.

[0096] In some embodiments, the analysis and inference model marks key frame blocks and other blocks separately to distinguish the types of blocks later.

[0097] S22. Mark the frame data in the key frame that corresponds to the key picture block as a key data packet.

[0098] S23. Mark the frame data in the keyframe that corresponds to the other picture blocks as other data packets.

[0099] S3. Encode the key data packet and the other data packets based on different resolutions to obtain a first data packet and a second data packet, wherein the resolution of the second data packet is lower than that of the first data packet.

[0100] Specifically, step S3 includes:

[0101] S31. Receive the first resolution selected by the user, and encode the key data packet based on the first resolution to obtain the first data packet.

[0102] S32. Obtain a second resolution with a resolution level lower than the first resolution, and encode the other data packets based on the second resolution to obtain a second data packet.

[0103] In some embodiments, if the user-selected fixed resolution has been received before step S1, the source video stream corresponding to the fixed resolution can be directly obtained. After obtaining the key data packets and other data packets in step S2, there is no need to encode the key data packets; they are directly marked as the first data packet for transmission. For other data packets, encoding is performed based on a resolution lower than the fixed resolution to obtain the second data packet for transmission. For example, if the user-selected fixed resolution is 8K, the 8K resolution source video stream can be directly obtained. After obtaining the key data packets and other data packets through the analysis and inference model, the key data packets are directly marked as the first data packet. Simultaneously, a 4K resolution, one level lower than 8K, is obtained, and other data packets are encoded based on the 4K resolution to obtain the second data packet.

[0104] In some embodiments, in step S1, a source video stream with a resolution between a first resolution and a second resolution can be obtained. After performing the above step S2 to obtain key data packets and other data packets, the key data packets are encoded based on the first resolution to obtain a first data packet, and the other data packets are encoded based on the second resolution to obtain a second data packet.

[0105] In some embodiments, step S3 can be implemented based on an AI encoding model. Specifically, the AI ​​encoding model identifies key frame blocks and other frame blocks based on an analytical inference model, extracts the frame data corresponding to the key frame blocks and other frame blocks from the corresponding key frames, and encodes the frame data corresponding to the key frame blocks based on a first resolution selected by the user to obtain a first data packet with enhanced clarity; simultaneously, it selects a second resolution lower than the first resolution to encode the frame data corresponding to other frame blocks to obtain a second data packet with weakened clarity; furthermore, the AI ​​encoding model sets different transmission tags for the first data packet and the second data packet respectively.

[0106] S4. Transmit the first data packet through a reliable transmission protocol to obtain a first data packet set.

[0107] In an optional implementation, step S4 is further included before:

[0108] S401. Mark the frame data in the first data packet that is adjacent to the other picture blocks and is within a preset range as edge frame data.

[0109] S402, Mark the frame data in the first data packet that is not adjacent to the other screen blocks or is outside the preset range as center frame data.

[0110] In some embodiments, the preset range is 100 frames, and the preset range depends on the specific screen range.

[0111] In some embodiments, frame data in other frame blocks that are adjacent to the key frame block and within a preset range are marked as adjacent frame data, and frame data in other frame blocks that are not adjacent to the key frame block or are outside the preset range are marked as non-center frame data. The adjacent frame data and the key data packet are encoded based on a first resolution to obtain adjacent frame data and a first data packet at the first resolution. Simultaneously, the non-center frame data is encoded based on a second resolution to obtain a second data packet. At this point, the adjacent frame data at the first resolution and the second data packet can be transmitted via an unreliable transmission protocol, while the first data packet is transmitted via a reliable transmission protocol. This method ensures the clarity of the key frame while also avoiding obvious dividing lines in the video frame.

[0112] Then step S4 includes:

[0113] S41. The center frame data is transmitted through the reliable transmission protocol, and the edge frame data is transmitted through the unreliable transmission protocol to obtain the first data packet set.

[0114] S5. Transmit the second data packet through an unreliable transmission protocol to obtain a second data packet set.

[0115] In some embodiments, the reliable transmission protocol is TCP (Transmission Control Protocol), and the unreliable transmission protocol is UDP (User Datagram Protocol), which allows for packet loss.

[0116] S6. The keyframes are spliced ​​together according to the first data packet set and the second data packet set to obtain the target video for playback.

[0117] Please refer to Figure 2 Embodiment two of the present invention is as follows:

[0118] A video transmission terminal 100 includes a memory 101, a processor 102, and a computer program stored in the memory 101 and running on the processor 102. When the processor 102 executes the computer program, it implements various steps in a video transmission method according to Embodiment 1.

[0119] In summary, the video transmission method and terminal provided by this invention identify key frame blocks based on an analytical inference model, thereby dividing frame data into key data packets and other data packets to distinguish between important and non-important data in the video. When the user selects a fixed resolution, key data packets that the user may be interested in are encoded at a fixed resolution to ensure video clarity, while other data packets that the user is not interested in are encoded at a lower resolution to reduce the amount of data transmitted. Simultaneously, a reliable transmission protocol is used when transmitting high-resolution data packets to avoid packet loss, thus ensuring data quality. An unreliable transmission protocol is used when transmitting the second data packet; even if the second data packet is lost, retransmission is unnecessary, thereby improving transmission speed. For frame data between different resolutions, high-resolution encoding and an unreliable transmission protocol are used to avoid obvious visual boundaries when stitching frame data of different resolutions. This invention can reduce data transmission volume and improve data transmission efficiency while ensuring video quality.

[0120] The above description is merely an embodiment of the present invention and does not limit the patent scope of the present invention. Any equivalent modifications made based on the content of the present invention specification and drawings, or direct or indirect applications in related technical fields, are similarly included within the patent protection scope of the present invention.

Claims

1. A method of video transmission, characterized by, The method comprises the following steps: acquiring a source video stream, and acquiring a set of image frames of continuous pictures from the source video stream as key frames; determining key data packets and other data packets in the key frames based on a preset analysis and inference model; encoding and processing the key data packets and the other data packets based on different resolutions respectively to obtain first data packets and second data packets, the resolution of the second data packets being lower than that of the first data packets; transmitting the first data packets through a reliable transmission protocol to obtain a first data packet set; transmitting the second data packets through an unreliable transmission protocol to obtain a second data packet set; splicing the key frames based on the first data packet set and the second data packet set to obtain a target video for playing; the step of determining the key data packets and the other data packets in the key frames based on the preset analysis and inference model comprises the following steps: identifying key picture blocks and other picture blocks in the key frames based on the preset analysis and inference model; marking frame data corresponding to the key picture blocks in the key frames as key data packets; marking frame data corresponding to the other picture blocks in the key frames as other data packets.

2. The video transmission method of claim 1, wherein, the analysis and inference model is trained based on the attention degree of a user to different picture positions in the key frames; the step of identifying the key picture blocks and the other picture blocks in the key frames based on the preset analysis and inference model comprises the following steps: acquiring a display picture of the key frames, and dividing the display picture into a plurality of picture blocks according to picture positions; performing target identification on the plurality of picture blocks in the key frames based on the preset analysis and inference model to obtain the key picture blocks and the other picture blocks of the key frames.

3. The video transmission method of claim 1, wherein, the step of encoding and processing the key data packets and the other data packets based on different resolutions respectively to obtain the first data packets and the second data packets comprises the following steps: receiving a first resolution selected by a user, and encoding and processing the key data packets based on the first resolution to obtain the first data packets; acquiring a second resolution lower than the first resolution in resolution level, and encoding and processing the other data packets based on the second resolution to obtain the second data packets.

4. The video transmission method of claim 1, wherein, the step of transmitting the first data packets through the reliable transmission protocol further comprises the following steps: marking frame data adjacent to the other picture blocks and within a preset range in the first data packets as edge frame data; marking frame data not adjacent to the other picture blocks or outside the preset range in the first data packets as center frame data; the step of transmitting the first data packets through the reliable transmission protocol to obtain the first data packet set comprises the following step: transmitting the center frame data through the reliable transmission protocol, and transmitting the edge frame data through an unreliable transmission protocol to obtain the first data packet set.

5. A video transmission terminal comprising a memory, a processor and a computer program stored on the memory and running on the processor, characterized in that, the processor executes the computer program to realize the following steps: acquiring a source video stream, and acquiring a set of image frames of continuous pictures from the source video stream as key frames; determining key data packets and other data packets in the key frames based on a preset analysis and inference model; The key data packet and the other data packet are encoded based on different resolutions to obtain a first data packet and a second data packet, and the resolution of the second data packet is lower than that of the first data packet; The first data packet is transmitted through a reliable transmission protocol to obtain a first data packet set; The second data packet is transmitted through an unreliable transmission protocol to obtain a second data packet set; The key frame is spliced based on the first data packet set and the second data packet set to obtain a target video for playing; The key data packet and the other data packet in the key frame are determined based on a preset analysis and inference model, which includes: The key frame is identified based on a preset analysis and inference model to obtain a key picture block and an other picture block; Frame data corresponding to the key picture block in the key frame is marked as a key data packet; Frame data corresponding to the other picture block in the key frame is marked as an other data packet.

6. A video transmission terminal according to claim 5, characterized in that The analysis and inference model is trained based on user attention to different picture positions in the key frame; The key picture block and the other picture block in the key frame are identified based on a preset analysis and inference model, which includes: An exhibition picture of the key frame is obtained, and the exhibition picture is divided into a plurality of picture blocks according to picture positions; The plurality of picture blocks in the key frame are target-identified based on a preset analysis and inference model to obtain a key picture block and an other picture block of the key frame.

7. A video transmission terminal according to claim 5, characterized in that The key data packet and the other data packet are encoded based on different resolutions to obtain a first data packet and a second data packet, which includes: A first resolution selected by a user is received, and the key data packet is encoded based on the first resolution to obtain a first data packet; A second resolution lower than the first resolution in resolution level is obtained, and the other data packet is encoded based on the second resolution to obtain a second data packet.

8. A video transmission terminal according to claim 5, characterized in that Before the first data packet is transmitted through a reliable transmission protocol, it further includes: Frame data in the first data packet adjacent to the other picture block and within a preset range is marked as edge frame data; Frame data in the first data packet not adjacent to the other picture block or outside the preset range is marked as center frame data; The first data packet is transmitted through a reliable transmission protocol to obtain a first data packet set, which includes: The center frame data is transmitted through the reliable transmission protocol, and the edge frame data is transmitted through an unreliable transmission protocol to obtain a first data packet set.

Images