Search

RU-2861389-C1 - SYSTEM FOR ESTIMATING VIDEO QUALITY DEGRADATION

RU2861389C1RU 2861389 C1RU2861389 C1RU 2861389C1RU-2861389-C1

Abstract

FIELD: video communication. SUBSTANCE: receiving an original video file from a storage device, wherein the original video file contains a sequence of numbered video frames. Sending the original video file to a network problem emulation module, which adds one or more network distortions for the video stream to an emulated communication channel for video transmission. Then, receiving a resulting video file that has passed through the emulated communication channel and storing it on the storage device. After that, a video quality degradation calculation module calculates a resulting video quality degradation estimate by comparing the original video file and the resulting video file stored on the storage device, wherein the estimate is composed of weighted parameters corresponding to the types of distortions that reduced the video quality during transmission over the emulated communication channel. EFFECT: ensuring increased accuracy of estimating video quality degradation when transmitting over problematic communication channels. 22 cl, 4 dwg

Inventors

  • Slavetskij Aleksandr Viktorovich

Dates

Publication Date
20260505
Application Date
20250731

Claims (20)

  1. 1. A computer system for assessing the reduction in quality of video transmitted over communication channels, comprising:
  2. storage device, and
  3. a data processing device operatively connected to a storage device, comprising at least a frame numbering module, a network problem emulation module and a video quality degradation estimation module,
  4. wherein the system is configured to perform the steps of:
  5. - the first receiving/transmitting module receives a source video file from a storage device, wherein the source video file contains a sequence of numbered video frames;
  6. - the first transmit/receive module sends the source video file to the network problem emulation module, which adds one or more network distortions to the video stream in the emulated video transmission channel;
  7. - the second receiving/transmitting module receives the resulting video file that has passed through the emulated communication channel and saves it on a storage device;
  8. - the video quality degradation calculation module calculates the resulting video quality degradation estimate by comparing the original video file and the resulting video file stored on the storage device, wherein the estimate is made up of weighted parameters corresponding to the types of distortions that reduced the video quality during transmission over the emulated communication channel.
  9. 2. The system according to paragraph 1, which is used for automatic testing of video communication.
  10. 3. The system of claim 1, wherein the possible types of communication channel distortions are pre-classified, and each type of distortion corresponds to a pre-set weighting coefficient, and the system allows the user to set and/or change the values of the weighting coefficients before starting the automatic testing.
  11. 4. The system of claim 1, configured to count the number of each type of distortion.
  12. 5. The system of claim 1, wherein the types of distortions being assessed are at least: (1) a reduction in the quality of video frames during encoding, characterized by increased granulation, a blurring effect, and frame fuzziness; (2) a reduction in the number of frames per second (FPS) during encoding of a video stream; (3) a visual effect of a long pause in playback, characterized by the repetition of the same frame; (4) skipping frame chains, leading to a long pause in video playback; (5) scattering of an image in a frame, characterized by frame allowance or incomplete frames.
  13. 6. The system according to paragraph 1, additionally configured with the ability to display a single resulting assessment to the user.
  14. 7. The system according to paragraph 1, additionally configured with the ability to display to the user a separate assessment for each of the types of distortions and/or the number of each of the identified types of distortions.
  15. 8. The system of claim 1, wherein the frame numbering module is configured to number each frame of the original video file and store the original video file containing the sequence of numbered video files.
  16. 9. The system of claim 8, wherein numbered markers representing small regions on the image are used to number each frame of the original video file, and the frame number is encoded into the color model components of each region.
  17. 10. The system according to claim 9, in which two color difference components (UV) of the YUV color model are used for frame numbering.
  18. 11. The system of claim 10, in which (5) scattering of an image in a frame is determined if, when comparing the original and resulting video file, the value of the frame number does not match for at least one number mark.
  19. 12. A method for assessing the reduction in video quality, performed by a computer system and comprising the steps of:
  20. - the first receiving/transmitting module receives a source video file from a storage device, wherein the source video file contains a sequence of numbered video frames;

Description

The present invention relates to video communication technologies, and more specifically to the automatic testing of hardware and software video conferencing systems by assessing the decline in the quality of a video stream transmitted over communication channels. LEVEL OF TECHNOLOGY Video calls are a common method of communication these days. Society increasingly relies on data networks for both work and communication. Video calls enable interaction between people who otherwise might not be able to meet in person for various reasons. Due to the growing popularity of video calls, maintaining the quality of video calls is becoming increasingly important. It's common knowledge that video conferencing is a direct visual connection between two or more remote parties over the internet, simulating a face-to-face meeting. Therefore, video conferencing is similar in its requirements to real-time communication with guaranteed information delivery. IP (Internet Protocol) is the primary network protocol used to transmit data over the internet. It addresses and routes data packets between devices, allowing them to communicate with each other. However, from the perspective of video conferencing developers, IP is an unstable environment for video conferencing. Video data transmission can be particularly susceptible to common network issues such as packet loss, delayed packets, and packet delays. If the required speed of information transfer from the sender's terminal to the recipient's terminal is not ensured without delays and loss of information, then organizing and using video conferencing will become inconvenient, technologically and algorithmically difficult, or even impossible. Video communication technologies use video encoders (video codecs) that compress frames for transmission in such a way that the frames of a video stream are interdependent. If even one frame is lost, the decoder will experience distortion that will negatively impact subsequent frames. The following IP protocols are commonly used to organize video conferences. 1) The connection-oriented TCP/IP transport protocol (TCP - Transmission Control Protocol) ensures the integrity and order of information, but on overloaded problematic communication channels it can introduce large delays. 2) The UDP/IP transport protocol (UDP - User Datagram Protocol) is essentially a superset of IP, adding port addressing. Therefore, it inherits all the properties of the IP protocol. The main advantage of UDP is the speed of packet transmission. Therefore, in most cases, UDP is used in video conferencing to ensure real-time communication. If UDP is not possible, a TCP connection is established. Moreover, it's not known in advance which protocol will be used for video communication. Therefore, the stacks for sending and receiving video streams on the sending and receiving terminals must be prepared for any negative properties of the combined properties of possible internet channels. The negative properties of packet transmission channels over IP networks include the fact that the IP network does not guarantee: (1) packet delivery, (2) the correct sequence and uniformity of packet arrival, (3) the integrity of transmitted packets. In addition to the above, the Internet cannot provide constant channel bandwidth and prevent delays in the transmission of packets from the sender to the recipient. To overcome the negative properties of IP networks for transmitted video information, modern hardware and software video conferencing systems use special protocols with feedback and well-known additional technologies. For example, for video streams, such well-known protocols and technologies are used as: RTP/RTCP (Realtime Transport Protocol / Realtime Transport Control Protocol), JitterBuffer, NACK (Negative Acknowledge) - re-requests for lost packets, FEC (Forward Error Correction), re-requests for restoring the image in frames, maintaining channel statistics and exchanging them between the sending terminal and the receiving terminal, etc. The most common technology for preventing the negative properties of IP networks is adapting the quality of the transmitted video stream to the properties of the channel. Some solutions for assessing the decline in video quality under the influence of negative IP network effects are known from the prior art: PSNR, SSIM, VMAF metrics, etc. Image evaluation metrics such as Peak Signal-to-Noise Ratio (PSNR) and Structure Similarity (SSIM) compare the original frame (image/picture) and the resulting frame (passed through the communication channel). However, in video communications, it is the video itself (i.e., the vectors of these frames over time) that must be evaluated, not individual frames. The above metrics cannot assess video quality during transmission, since some frames may be dropped during channel adaptation (for example, by reducing the number of frames per second), or, conversely, some frames may be duplicated. The well-known VMAF (Vid