Reliability-Aware Prediction via Uncertainty Learning for Person Image Retrieval

TL;DR

This paper introduces an Uncertainty-Aware Learning method for person image retrieval that estimates prediction reliability by modeling data and model uncertainties, improving ranking accuracy and confidence assessment.

Contribution

It proposes a novel unified framework that jointly learns data and model uncertainties without sampling, enhancing reliability-aware predictions in person re-identification.

Findings

Effective reliability assessment demonstrated in experiments.

Superior performance on three benchmarks.

Improved confidence estimation for predictions.

Abstract

Current person image retrieval methods have achieved great improvements in accuracy metrics. However, they rarely describe the reliability of the prediction. In this paper, we propose an Uncertainty-Aware Learning (UAL) method to remedy this issue. UAL aims at providing reliability-aware predictions by considering data uncertainty and model uncertainty simultaneously. Data uncertainty captures the ``noise" inherent in the sample, while model uncertainty depicts the model's confidence in the sample's prediction. Specifically, in UAL, (1) we propose a sampling-free data uncertainty learning method to adaptively assign weights to different samples during training, down-weighting the low-quality ambiguous samples. (2) we leverage the Bayesian framework to model the model uncertainty by assuming the parameters of the network follow a Bernoulli distribution. (3) the data uncertainty and the model uncertainty are jointly learned in a unified network, and they serve as two fundamental criteria for the reliability assessment: if a probe is high-quality (low data uncertainty) and the model is confident in the prediction of the probe (low model uncertainty), the final ranking will be assessed as reliable. Experiments under the risk-controlled settings and the multi-query settings show the proposed reliability assessment is effective. Our method also shows superior performance on three challenging benchmarks under the vanilla single query settings.