An Inter-observer consistent deep adversarial training for visual scanpath prediction

TL;DR

This paper introduces a novel adversarial training method for visual scanpath prediction that enhances inter-observer consistency using a lightweight neural network, improving state-of-the-art performance.

Contribution

It presents an inter-observer consistent adversarial training framework for scanpath prediction, addressing the stochastic and subjective nature of visual attention.

Findings

Outperforms existing methods in accuracy

Maintains inter-observer consistency

Effective with a lightweight neural network

Abstract

The visual scanpath is a sequence of points through which the human gaze moves while exploring a scene. It represents the fundamental concepts upon which visual attention research is based. As a result, the ability to predict them has emerged as an important task in recent years. In this paper, we propose an inter-observer consistent adversarial training approach for scanpath prediction through a lightweight deep neural network. The adversarial method employs a discriminative neural network as a dynamic loss that is better suited to model the natural stochastic phenomenon while maintaining consistency between the distributions related to the subjective nature of scanpaths traversed by different observers. Through extensive testing, we show the competitiveness of our approach in regard to state-of-the-art methods.