Weak Hypothesis Generation for Face Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing face detection techniques are time-consuming due to the large number of possible filters and the need for redundant arithmetic operations, especially when dealing with variable-sized search windows and high-speed learning of weak hypotheses.
Innovation Solution
A weak hypothesis generation apparatus that selectively generates high-performance weak hypotheses by modifying existing ones, reducing the number of arithmetic operations required while maintaining accuracy, using data weighting and ensemble learning to improve discrimination speed and efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If all possible filters are used for face detection, then detection accuracy is improved, but processing time increases significantly
Solution Approach 1:
The patent segments the large set of all possible filters into two groups: a first set of filters that are actually applied for detection, and a second set of filters that are generated but not applied. This segmentation allows the system to maintain high detection accuracy by using the first set while reducing processing time by excluding the second set, resolving the contradiction between accuracy and speed.
Solution Approach 2:
The patent applies partial action by using only a subset of all possible filters (the first set) rather than all filters (the second set). This partial application of filters maintains sufficient detection accuracy while significantly reducing the computational load and processing time, thus resolving the contradiction between precision and time loss.
2Measurement precision
If redundant arithmetic operations are performed for high-speed learning, then learning accuracy is improved, but computational efficiency deteriorates
Solution Approach 1:
The patent extracts and removes redundant arithmetic operations from the learning process. By identifying and excluding unnecessary computations, the system maintains learning accuracy through essential operations while improving computational efficiency by eliminating redundant calculations, thus resolving the contradiction between precision and productivity.
Solution Approach 2:
The patent applies partial action by performing only the necessary arithmetic operations required for high-speed learning rather than all possible operations. This selective execution maintains learning accuracy while significantly improving computational efficiency by avoiding redundant calculations, resolving the contradiction between precision and productivity.
3Adaptability or versatility
If variable-sized search windows are processed, then detection versatility is improved, but processing speed decreases
Solution Approach 1:
The patent segments the processing of variable-sized search windows into manageable portions by using the first set of filters for actual detection and the second set for generation only. This segmentation allows the system to handle different search window sizes flexibly while maintaining high processing speed by not performing redundant operations on all possible filter combinations, thus resolving the contradiction between versatility and speed.
Data Source
AI summary
A facial expression recognition system that uses a face detection apparatus realizing efficient learning and high-speed detection processing based on ensemble learning when detecting an area representing a detection target and that is robust against shifts of face position included in images and capable of highly accurate expression recognition, and a learning method for the system, are provided. When learning data to be used by the face detection apparatus by Adaboost, processing to select high-performance weak hypotheses from all weak hypotheses, then generate new weak hypotheses from these high-performance weak hypotheses on the basis of statistical characteristics, and select one weak hypothesis having the highest discrimination performance from these weak hypotheses, is repeated to sequentially generate a weak hypothesis, and a final hypothesis is thus acquired. In detection, using an abort threshold value that has been learned in advance, whether provided data can be obviously judged as a non-face is determined every time one weak hypothesis outputs the result of discrimination. If it can be judged so, processing is aborted. A predetermined Gabor filter is selected from the detected face image by an Adaboost technique, and a support vector for only a feature quantity extracted by the selected filter is learned, thus performing expression recognition.