Size-invariance Matters: Rethinking Metrics and Losses for Imbalanced Multi-object Salient Object Detection

TL;DR

This paper highlights the importance of size-invariant evaluation metrics in multi-object salient detection, proposing a new approach that evaluates objects individually to improve detection across diverse sizes.

Contribution

It introduces size-invariant metrics and a novel evaluation framework that enhances multi-object detection performance by addressing size bias in existing metrics.

Findings

Improved detection of small and large objects across datasets.

New metrics better reflect true saliency regardless of object size.

The proposed method outperforms existing approaches in experiments.

Abstract

This paper explores the size-invariance of evaluation metrics in Salient Object Detection (SOD), especially when multiple targets of diverse sizes co-exist in the same image. We observe that current metrics are size-sensitive, where larger objects are focused, and smaller ones tend to be ignored. We argue that the evaluation should be size-invariant because bias based on size is unjustified without additional semantic information. In pursuit of this, we propose a generic approach that evaluates each salient object separately and then combines the results, effectively alleviating the imbalance. We further develop an optimization framework tailored to this goal, achieving considerable improvements in detecting objects of different sizes. Theoretically, we provide evidence supporting the validity of our new metrics and present the generalization analysis of SOD. Extensive experiments demonstrate the effectiveness of our method. The code is available at https://github.com/Ferry-Li/SI-SOD.