Adaptive Context-Aided Human Classification Using Dynamic Formula Selection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing human recognition techniques in images face challenges due to variations in lighting, pose, image quality, facial expressions, and aging, as well as the assumption of independent facial and contextual features, which hampers accurate identification and classification.
Innovation Solution
An adaptive context-aided human classification method that integrates face and clothes recognition data using various formulas based on the availability of features, employing linear logistic regression to estimate probabilities of identity similarity, even when face or clothes recognition results are missing.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If facial features and contextual features are assumed to be independent for human identification, then the identification method is simpler, but the accuracy of characterizing people is reduced
Solution Approach 1:
The patent merges facial features and contextual features into a unified probabilistic framework where both features contribute to the overall identification probability. Instead of treating them independently, the system combines face recognition scores and clothes recognition scores using probability formulas that account for their joint contribution to identifying whether two images show the same person.
2Productivity
If face and clothes recognition data are integrated using a fixed formula, then the processing is faster, but the accuracy is reduced when features are unavailable
Solution Approach 1:
The patent implements a dynamic formula selection mechanism that adapts the integration approach based on data availability. The system evaluates whether face scores, clothes scores, or both are available and selects the appropriate probability formula accordingly. This dynamic adaptation ensures accurate results regardless of which features are present or missing in the input images.
Solution Approach 2:
The system changes the parameters of the probability calculation based on feature availability. When both face and clothes scores are available, a combined probability formula is used. When only one type of score is available, the system switches to a formula that relies solely on that score type, thereby maintaining accuracy across varying data conditions.
3Device complexity
If only face recognition is used for identification, then the method is simpler, but the reliability is reduced under varying imaging conditions
Solution Approach 1:
The patent creates a multi-functional recognition system that can operate using face recognition, clothes recognition, or both together depending on the situation. The unified probabilistic framework allows the system to universally handle different imaging conditions by selecting the appropriate recognition modality, thereby improving reliability across diverse scenarios such as poor lighting, occlusion, or pose variations.
Data Source
AI summary
A method and an apparatus process digital images. The method according to one embodiment accesses digital data representing a plurality of digital images including a plurality of persons; performs face recognition to determine first scores relating to similarity between faces of the plurality of persons; performs clothes recognition to determine second scores relating to similarity between clothes of the plurality of persons; provides a plurality of formulas for estimating a probability of a face from the faces and a clothes from the clothes to belong to a person from the plurality of persons, wherein at least one formula of the plurality of formulas utilizes a first score and a second score, and at least one formula of the plurality of formulas utilizes only one score of a first score and a second score; and selects a formula from the plurality of formulas based on availability of a first score from the first scores for two persons from the plurality of persons, and availability of a second score from the second scores for the two persons, the selected formula estimating a probability relating to similarity of identities of the two persons.