Likelihood ratio for a binary Bayesian classifier under a noise-exclusion model

2026-01-12

AI Summary

This study addresses holistic visual search tasks under noisy conditions by proposing a parameter-efficient binary Bayesian classifier. Grounded in ideal observer theory, the method establishes a decision threshold based on minimally extractable image features and incorporates a noise-rejection mechanism alongside feature constraints, substantially reducing model degrees of freedom and computational complexity. The resulting statistical model maintains high discriminative performance while enabling practical applications in system evaluation and feature selection for domains such as medical image analysis and security-related object recognition. By offering a theoretically principled yet computationally tractable approach, this work provides an effective new tool for imaging system optimization and benchmarking in computer vision.

Technology Category

• Computer Vision: Image and Video Retrieval
• Search and Optimization: Learning to Search
• Machine Learning: Bayesian Learning

Application Category

• User Modeling, Personalization and Recommendation: ML for personalized search and recommendations
• Search and Retrieval-Augmented AI: Web evaluation methodologies and metrics
• Graph Algorithms and Modeling for the Web: Algorithms and analysis for incomplete, noisy, or partially observed Web-related graphs

Abstract

We develop a new statistical ideal observer model that performs holistic visual search (or gist) processing in part by placing thresholds on minimum extractable image features. In this model, the ideal observer reduces the number of free parameters thereby shrinking down the system. The applications of this novel framework is in medical image perception (for optimizing imaging systems and algorithms), computer vision, benchmarking performance and enabling feature selection/evaluations. Other applications are in target detection and recognition in defense/security as well as evaluating sensors and detectors.

Problem

• likelihood ratio
• binary Bayesian classifier
• noise-exclusion model
• ideal observer
• feature extraction

Innovation

• ideal observer model
• noise-exclusion
• binary Bayesian classifier
• feature thresholding
• holistic visual search