Face Recognition Using Adaboost Trait Codes
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current face recognition and synthesis methods are limited by the subjective selection of facial features, leading to inaccurate and time-consuming comparisons, and lack the necessary discriminating capability for effective recognition and synthesis.
Innovation Solution
The method employs Facial Trait Code, generated by classifying human faces into N most discriminating local features using the Adaboost algorithm, constructing a code book, and encoding principal traits to produce a serial code for efficient comparison and synthesis, utilizing distances like Hamming, Euclidean, or Bhattacharyya distance for accurate matching and fault tolerance.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If subjective human instinct is used to select facial features, then the method is simple to implement, but the discriminating capability and recognition accuracy are insufficient
Solution Approach 1:
The patent replaces the mechanical/manual process of subjective feature selection with an automated computer-based system that uses algorithms (such as Adaboost) to objectively select and extract facial features. This substitution eliminates human subjectivity and instinct-based selection, providing consistent, reproducible, and optimized feature selection that improves recognition accuracy while maintaining implementation feasibility through automated processing.
2Productivity
If direct vector conversion of human face is used, then the dimension is reduced, but the slight difference on face leads to gigantic divergence between vectors
Solution Approach 1:
The patent divides the face image into multiple local feature regions (such as eyes, nose, mouth, cheeks) and extracts features independently from each segment. Instead of converting the entire face into a single high-dimensional vector, the system processes segmented feature regions separately, which reduces the impact of local variations and prevents slight differences from causing gigantic divergence in the overall comparison.
Solution Approach 2:
The patent applies different processing and comparison strategies to different facial feature regions based on their local characteristics. Each facial segment is evaluated with appropriate weight and processing method, allowing the system to focus on discriminative local features while being tolerant to minor variations in non-critical regions, thereby improving vector comparison accuracy.
3Measurement precision
If predefined facial components are used for synthesis, then the components are easy to select, but they lack representative discrimination property
Solution Approach 1:
The patent replaces manual, instinct-based selection of facial components for synthesis with an automated system that uses the same objective algorithms (such as Adaboost) to identify and select the most discriminative feature regions. This automated component selection process ensures that the synthesized facial features have strong discrimination properties while the system remains easy to implement through algorithmic automation.
Data Source
AI summary
A method of recognizing an object in an image is provided, the method comprises the following steps. The image having the object is provided, and principal traits of the object are encoded in order to generate a first trait code. The first trait code is compared with data stored in a database so as to obtain a plurality of differences. A minimum of the plurality of differences is found. This method can be applied to synthesize human faces.