Video Stream Transport via Visual Sensitivity Assessment
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current packet discard mechanisms in network congestion do not consider the varying visual impact of lost IPTV packets, leading to suboptimal user experience, particularly in real-time video streaming, and are limited to specific coding and encapsulation mechanisms like H.264 and RTP.
Innovation Solution
A method and system that assesses video data packets to determine their visual sensitivity and adds this information to the packets, allowing for intelligent handling and selective discarding based on the impact of loss on the terminal screen, supporting multiple coding mechanisms such as H.264, MPEG-2, and encapsulation mechanisms like RTP and MPEG2-TS.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If packets are discarded randomly according to traditional IP QoS algorithms, then network congestion is relieved, but user visual experience deteriorates due to loss of visually sensitive packets
Solution Approach 1:
The patent applies local quality by differentiating packet discard strategies based on visual sensitivity. Instead of uniform random discard, the system identifies packets with different visual importance (using markers like NRI in H.264 or reference frame indicators) and applies selective discard policies. Visually less important packets are discarded first, while visually critical packets are preserved, thereby maintaining visual quality while managing congestion.
Solution Approach 2:
The patent changes the discard parameter from random probability-based selection to visual sensitivity-based selection. By introducing visual sensitivity as a new discarding parameter and using packet markers to indicate visual importance, the system transforms the discard mechanism into a quality-aware process that considers the impact of each packet on visual output rather than treating all packets equally.
2Productivity
If packet discard is based on frame type (I, P, B frames) as in H.264 NRI field, then congestion is managed, but the solution is limited to H.264 coding and RTP encapsulation only
Solution Approach 1:
The patent achieves universality by creating a marker-based identification mechanism that can work across multiple video coding formats (H.264, MPEG-2, VC-1) and encapsulation types (RTP, MPEG2-TS). Instead of hardcoding format-specific discard logic, the system uses universal markers (such as NRI field in H.264, or reference frame indicators in MPEG-2) that can be detected and processed by network devices regardless of the underlying coding or encapsulation scheme, making the solution adaptable to diverse video stream formats.
3Device complexity
If entire frames are discarded based on NRI value, then processing is simplified, but packet loss impact on user experience is worsened because not all packets in a frame have equal visual importance
Solution Approach 1:
The patent applies segmentation by dividing the frame into individual packets and assigning different discard priorities to packets within the same frame based on their visual importance. Instead of treating all packets in a frame uniformly, the system identifies visually critical packets (such as those containing important macroblocks or reference information) and protects them from discard, while allowing less critical packets to be discarded during congestion, thereby segmenting the discard decision at the packet level rather than the frame level.
Data Source
AI summary
A method, an apparatus, and a system for transporting video streams are disclosed herein to handle video streams of different coding mechanisms and different encapsulation mechanisms intelligently according to visual sensitivity. The method includes: assessing a video data packet to gain visual sensitivity identifier information of the video data packet; adding the visual sensitivity identifier information into the video data packet; and handling the video data packet according to the visual sensitivity identifier information while transporting the video data packet. The embodiments of the present invention are applicable to transporting of network video streams.