Energy Score-based Pseudo-Label Filtering and Adaptive Loss for Imbalanced Semi-supervised SAR target recognition

TL;DR

This paper introduces a novel semi-supervised SAR target recognition method that uses energy scores and adaptive loss functions to improve accuracy in highly imbalanced datasets, addressing class imbalance issues effectively.

Contribution

The paper proposes a dynamic energy score-based pseudo-label selection and adaptive loss functions specifically designed for imbalanced semi-supervised SAR ATR tasks, enhancing recognition accuracy.

Findings

Effective pseudo-label selection with energy scores improves reliability.

Adaptive loss functions mitigate class imbalance effects.

High-precision recognition achieved on imbalanced SAR datasets.

Abstract

Automatic target recognition (ATR) is an important use case for synthetic aperture radar (SAR) image interpretation. Recent years have seen significant advancements in SAR ATR technology based on semi-supervised learning. However, existing semi-supervised SAR ATR algorithms show low recognition accuracy in the case of class imbalance. This work offers a non-balanced semi-supervised SAR target recognition approach using dynamic energy scores and adaptive loss. First, an energy score-based method is developed to dynamically select unlabeled samples near to the training distribution as pseudo-labels during training, assuring pseudo-label reliability in long-tailed distribution circumstances. Secondly, loss functions suitable for class imbalances are proposed, including adaptive margin perception loss and adaptive hard triplet loss, the former offsets inter-class confusion of classifiers, alleviating the imbalance issue inherent in pseudo-label generation. The latter effectively tackles the model's preference for the majority class by focusing on complex difficult samples during training. Experimental results on extremely imbalanced SAR datasets demonstrate that the proposed method performs well under the dual constraints of scarce labels and data imbalance, effectively overcoming the model bias caused by data imbalance and achieving high-precision target recognition.