Changes in Visual Attention Patterns for Detection Tasks due to Dependencies on Signal and Background Spatial Frequencies

2026-01-13

arXiv.org

AI Summary

This study investigates how the spatial frequency characteristics of target signals and anatomical background jointly influence visual attention and lesion detection accuracy in digital breast tomosynthesis (DBT). Using simulated DBT images generated from digital breast phantoms (Bakic/XCAT), the research integrates eye-tracking with human observer experiments to analyze fixation behavior and detection performance for various lesions—such as spiculated masses—under varying levels of background complexity. The findings reveal a critical interaction between signal morphology and anatomical noise, demonstrating that detection errors primarily arise from post-perceptual processing and decision biases. Notably, this work provides the first evidence that visual attention dynamically modulates fixation duration based on the spatial frequency properties of both signal and background, highlighting a synergistic constraint mechanism whereby local signal features and global background complexity jointly govern detection performance.

Technology Category

• Computer Vision: Medical and Biological Imaging
• Knowledge Representation and Reasoning: Geometric, Spatial, and Temporal Reasoning
• Data Mining & Knowledge Management: Mining of Visual, Multimedia & Multimodal Data

Application Category

• Economics, Online Markets and Human Computation: Data quality aspects of human-annotated datasets
• Graph Algorithms and Modeling for the Web: Representation, reconstruction, and subgraph or motif discovery in Web-related graphs
• Web Mining and Content Analysis: Robustness and generalizability of Web mining methods

Abstract

We aim to investigate the impact of image and signal properties on visual attention mechanisms during a signal detection task in digital images. The application of insight yielded from this work spans many areas of digital imaging where signal or pattern recognition is involved in complex heterogenous background. We will use simulated tomographic breast images as the platform to investigate this question. While radiologists are highly effective at analyzing medical images to detect and diagnose diseases, misdiagnosis still occurs. These errors can stem from multiple sources such as the inherent properties of medical images and the accompanying signal characteristics play a pivotal role in visual search strategy while locating and identifying the disease. We selected digital breast tomosynthesis (DBT) images as a sample medical images with different breast densities and structures using digital breast phantoms (Bakic and XCAT). Two types of lesions (with distinct spatial frequency properties) were randomly inserted in the phantoms during projections to generate abnormal cases. Six non-radiologist human observers participated in observer study designed for a locating and detection of an 3-mm sphere lesion and 6-mm spicule lesion in reconstructed in-plane DBT slices. We collected eye-gaze data to estimate gaze metrics and to examine differences in visual attention mechanisms. Gaze analysis revealed that diagnostic response times were significantly longer in Bakic phantoms and high-density tissue backgrounds compared to XCAT breast phantoms (p ¡ 0.05) and lower-density backgrounds (p ¡ 0.01) respectively, indicating increased perceptual difficulty in anatomically complex scenes. Observers made longer fixations on spiculated lesions compared to spherical lesions (p ¡ 0.01), suggesting that lesion morphology modulates visual attention allocation. Across all attention stages—search, recognition, and decision—the small spherical lesion required higher contrast for successful detection, especially in the Bakic background where anatomical noise severely reduced its visibility. Error analysis showed that the majority of misdiagnosed cases (25%) occurred when the lesion was not visible, followed by decision-stage errors (11%), where the signal was seen but incorrectly judged. Conclusions: Diagnostic performance in complex visual environments is strongly constrained by later perceptual stages, with decision failures accounting for the largest proportion of errors. Lesion detectability is jointly influenced by both target morphology and background complexity, revealing a critical interaction between local signal features and global anatomical noise. Increased fixation duration on spiculated lesions suggests that visual attention is differentially engaged depending on lesion structure, potentially aiding recognition. These findings highlight the importance of perceptually informed design and training of computer aided diagnosis systems.

Problem

• visual attention
• signal detection
• spatial frequency
• background complexity
• digital breast tomosynthesis

Innovation

• visual attention
• spatial frequency
• digital breast tomosynthesis
• eye-tracking
• signal detection