Biologically Inspired Hierarchical Model for Feature Extraction and Localization

TL;DR

This paper introduces a biologically inspired hierarchical model for feature extraction and localization in computer vision, mimicking human visual attention and coarse-to-fine search mechanisms to improve efficiency and accuracy.

Contribution

The paper proposes a novel hierarchical searching model that simulates human visual attention and coarse-to-fine search, enhancing feature localization in computer vision tasks.

Findings

Effective in localizing features with improved efficiency

Hierarchical model mimics biological visual attention mechanisms

Experimental results validate the model's accuracy and efficiency

Abstract

Feature extraction and matching are among central problems of computer vision. It is inefficent to search features over all locations and scales. Neurophysiological evidence shows that to locate objects in a digital image the human visual system employs visual attention to a specific object while ignoring others. The brain also has a mechanism to search from coarse to fine. In this paper, we present a feature extractor and an associated hierarchical searching model to simulate such processes. With the hierarchical representation of the object, coarse scanning is done through the matching of the larger scale and precise localization is conducted through the matching of the smaller scale. Experimental results justify the proposed model in its effectiveness and efficiency to localize features.