De-interlacing Video Comb Detection and Field Fusion
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video deinterlacing methods fail to ensure the absence of loss or combing in high-frequency signals in the vertical direction of deinterlaced images from interlaced videos.
Innovation Solution
A video processing method that acquires interlaced frame images, performs temporal and spatial deinterlacing, and conducts directional combing detection to determine if combing exists, then selects either the temporal or spatial virtual field image to fuse with the other field image, ensuring no loss or combing in the new image.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Speed
If temporal deinterlacing is performed to obtain temporal virtual field image, then temporal resolution is improved and picture smoothness is enhanced, but combing artifacts occur in high-frequency vertical signals
Solution Approach 1:
The patent performs preliminary combing detection on the temporal virtual field image before final image generation. By detecting combing artifacts in advance along candidate directions and comparing with reference field images, the system identifies regions with combing artifacts and prevents their propagation to the final output image, thus resolving the contradiction between temporal resolution improvement and combing artifact suppression.
2Object-affected harmful factors
If spatial deinterlacing is performed to obtain spatial virtual field image, then combing artifacts are reduced, but temporal resolution decreases and picture smoothness deteriorates
Solution Approach 1:
The patent applies different deinterlacing strategies to different regions of the image based on local characteristics. By performing directional combing detection and identifying regions with combing artifacts, the system selectively processes only those regions that require spatial deinterlacing, while preserving temporal virtual field image quality in regions without combing artifacts, thus maintaining overall temporal resolution while reducing combing artifacts locally.
Solution Approach 2:
The patent dynamically switches between temporal and spatial deinterlacing approaches based on real-time combing detection results. The system adapts its processing strategy by comparing candidate directions with reference images and adjusting the deinterlacing method according to the detected combing characteristics, optimizing the balance between temporal resolution and combing artifact suppression for each specific region and scene condition.
3Measurement precision
If directional combing detection is performed along candidate directions, then detection precision is improved, but device complexity increases
Solution Approach 1:
The patent performs combing detection along a limited set of candidate directions rather than exhaustively searching all possible directions. By identifying the most likely combing directions based on image characteristics and performing detection only along these candidate directions, the system achieves sufficient detection precision while significantly reducing computational complexity compared to omnidirectional detection approaches.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
The present invention discloses a video processing method and apparatus, which relate to the field of multimedia, and can ensure that there is neither loss nor combing in high-frequency signals in a perpendicular direction in a new image obtained after deinterlacing is performed on an interlaced frame image in an interlaced video. A specific solution is: acquiring, by an video processing apparatus, two images of an interlaced frame image in an interlaced video; performing temporal deinterlacing on either field image in the two images to obtain a temporal virtual field image, and performing spatial deinterlacing on the either field image to obtain a spatial virtual field image; performing directional combing detection on the temporal virtual field image along a preset candidate combing detection direction with reference to the other field image in the interlaced frame image, to determine whether combing exists in the temporal virtual field image; and according to this, selecting the temporal virtual field image or the spatial virtual field image to be fused with the other field image to generate a new image. The present invention is applicable to video processing.