Video Frame Splicing with Adaptive Key Frames for Blur Reduction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The image quality in videos generated by conventional encoding methods is compromised due to the use of a fixed IDR frame, leading to blurred transitions between segments of multimedia material.
Innovation Solution
Decoding multiple segments of multimedia material using different decoders and generating a key frame based on the renderer when adjacent frames are decoded by different decoders, allowing the encoder to perform encoding operations accordingly.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If a fixed IDR frame is used for encoding subsequent frames, then the encoding process is simplified and efficient, but the image quality deteriorates due to large differences between adjacent segments of multimedia material
Solution Approach 1:
The patent applies dynamics by making the IDR frame selection adaptive rather than fixed. The system dynamically determines whether to generate an IDR frame at each splicing position based on comparing image differences between adjacent frames. When the difference exceeds a threshold, an IDR frame is generated; otherwise, the previous frame is used. This dynamic adjustment resolves the contradiction by maintaining encoding efficiency while improving image quality at critical transition points.
2Ease of operation
If an IDR frame from the previous segment is used for encoding the subsequent segment, then the encoding operation is straightforward, but the transition between segments suffers from image quality lag and blurring
Solution Approach 1:
The patent implements feedback by calculating the image difference between adjacent frames at splicing positions and using this feedback to determine IDR frame generation. The system computes the absolute difference between pixel values of corresponding positions in adjacent frames, compares it against a threshold, and adjusts encoding accordingly. This feedback mechanism ensures reliable transitions by adapting encoding parameters to the actual content differences, resolving the contradiction between operational simplicity and transition quality.
3Productivity
If conventional encoding with fixed IDR frames is used, then processing speed is maintained, but image quality at splicing positions deteriorates due to large segment differences
Solution Approach 1:
The patent applies local quality by applying different encoding strategies to different regions of the video timeline. At splicing positions where image differences exceed the threshold, full IDR frame encoding is applied to ensure high quality. In regions with small differences, predictive encoding is used to maintain processing speed. This localized adaptation resolves the contradiction by optimizing both quality and speed for different segments of the video based on actual content characteristics.
Data Source
AI summary
Embodiments of the present disclosure provide a video processing method, apparatus and system, an electronic device, a storage medium, a computer program product, and a computer program. The method comprises: acquiring multimedia information, the multimedia information comprising multiple segments of multimedia material, and decoding the multiple segments of multimedia material on the basis of different decoders to obtain video frames of the segments of multimedia material, wherein one decoder decodes non-adjacent segments of multimedia material, and one segment of multimedia material corresponds to at least one video frame; and in response to two adjacent video frames received by a renderer being decoded by different decoders, generating a key frame on the basis of the renderer, outputting the key frame to an encoder, and on the basis of an encoding operation of the encoder according to the key frame, generating a video of the multimedia information.