Image Recognition Method for Video Frame Type Classification
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video processing technologies face challenges in accurately recognizing frame types in image frame sequences, leading to inefficiencies in frame rate conversion and video playback, particularly due to the presence of overlapped and repeated frame images.
Innovation Solution
An image recognition method that predicts a pixel sequence of a frame image based on neighboring frames, calculates a pixel error, and recognizes the frame type, effectively improving the accuracy and stability of frame type recognition.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If variable frame rate algorithm is used to convert frame rates, then frame rate unification is achieved, but overlapped and repeated frame images are generated
Solution Approach 1:
The patent segments the frame identification process into multiple stages: first identifying clear frames as reference points, then using these reference frames to identify surrounding frames through comparison. This segmentation approach allows accurate identification despite the presence of overlapped and repeated frames generated by variable frame rate conversion.
Solution Approach 2:
The patent introduces reference frames as intermediaries in the frame identification process. These reference frames serve as mediators to compare and identify the frame types of surrounding frames, enabling accurate frame type recognition even when direct comparison is difficult due to overlapped or repeated content.
2Loss of information
If frame type recognition is performed on all frames, then complete frame type information is obtained, but processing time and computational complexity increase
Solution Approach 1:
The patent performs preliminary identification of clear frames as reference frames before using them to identify surrounding frames. This preliminary action reduces the overall processing time by avoiding redundant comparisons, as reference frames are identified once and then used to efficiently identify multiple surrounding frames.
Solution Approach 2:
The patent performs frame type identification on frames surrounding reference frames with higher scrutiny, while using the reference frames themselves as anchors. This partial action approach focuses computational resources on critical frames, achieving complete frame type information without uniformly processing every frame at maximum detail.
3Speed
If overlapped and repeated frames are retained in the sequence, then frame rate requirements are met, but video playback quality deteriorates
Solution Approach 1:
The patent extracts and identifies overlapped and repeated frames through frame type recognition, enabling their subsequent removal or special handling. By taking out these problematic frames from the general sequence and treating them differently, the system maintains frame rate requirements while improving playback quality.
Solution Approach 2:
The patent changes the parameter of frame selection by identifying and separating frames of different types (clear frames vs. overlapped/repeated frames). This parameter change allows the system to selectively use appropriate frames for playback, maintaining frame rate while improving quality by avoiding problematic overlapped and repeated frames.
Data Source
AI summary
The disclosure provides an image recognition method and apparatus. The image recognition method may include obtaining a first frame image of which a frame type is known, and a second frame image and a third frame image of which frame types are unknown from an image frame sequence. The method may further include predicting a pixel sequence of the second frame image according to the first frame image and the third frame image. The method may further include calculating a pixel error according to a second pixel sequence of the second frame image and the predicted pixel sequence. The second pixel sequence includes pixel values of the pixels in the second frame image. The method may further include recognizing a frame type of the second frame image or the third frame image according to the pixel error.