Predicting the Category and Attributes of Visual Search Targets Using Deep Gaze Pooling

TL;DR

This paper introduces a novel gaze pooling layer for CNNs that predicts categories and attributes of visual search targets from eye gaze data, improving accuracy without extensive retraining.

Contribution

The authors propose a Gaze Pooling Layer that integrates gaze data into CNNs as an attention mechanism, enabling effective search target prediction with minimal retraining.

Findings

Effective gaze-based search target prediction demonstrated

Gaze pooling layer improves recognition accuracy

Method works with pre-trained CNNs without retraining

Abstract

Predicting the target of visual search from eye fixation (gaze) data is a challenging problem with many applications in human-computer interaction. In contrast to previous work that has focused on individual instances as a search target, we propose the first approach to predict categories and attributes of search targets based on gaze data. However, state of the art models for categorical recognition, in general, require large amounts of training data, which is prohibitive for gaze data. To address this challenge, we propose a novel Gaze Pooling Layer that integrates gaze information into CNN-based architectures as an attention mechanism - incorporating both spatial and temporal aspects of human gaze behavior. We show that our approach is effective even when the gaze pooling layer is added to an already trained CNN, thus eliminating the need for expensive joint data collection of visual and gaze data. We propose an experimental setup and data set and demonstrate the effectiveness of our method for search target prediction based on gaze behavior. We further study how to integrate temporal and spatial gaze information most effectively, and indicate directions for future research in the gaze-based prediction of mental states.