Scene Clustering Based Pseudo-labeling Strategy for Multi-modal Aerial View Object Classification

TL;DR

This paper introduces SCP-Label, a scene clustering pseudo-labeling strategy that significantly improves multi-modal aerial view object classification accuracy by leveraging scene aggregation properties, outperforming baselines and winning a challenge.

Contribution

The paper proposes a novel scene clustering based pseudo-labeling method (SCP-Label) that enhances classification accuracy in MAVOC by exploiting scene aggregation characteristics.

Findings

Achieved +20.57% accuracy on Track 1 (SAR)

Achieved +31.86% accuracy on Track 2 (SAR+EO)

Won the CVPR 2022 MAVOC Challenge in both tracks

Abstract

Multi-modal aerial view object classification (MAVOC) in Automatic target recognition (ATR), although an important and challenging problem, has been under studied. This paper firstly finds that fine-grained data, class imbalance and various shooting conditions preclude the representational ability of general image classification. Moreover, the MAVOC dataset has scene aggregation characteristics. By exploiting these properties, we propose Scene Clustering Based Pseudo-labeling Strategy (SCP-Label), a simple yet effective method to employ in post-processing. The SCP-Label brings greater accuracy by assigning the same label to objects within the same scene while also mitigating bias and confusion with model ensembles. Its performance surpasses the official baseline by a large margin of +20.57% Accuracy on Track 1 (SAR), and +31.86% Accuracy on Track 2 (SAR+EO), demonstrating the potential of SCP-Label as post-processing. Finally, we win the championship both on Track1 and Track2 in the CVPR 2022 Perception Beyond the Visible Spectrum (PBVS) Workshop MAVOC Challenge. Our code is available at https://github.com/HowieChangchn/SCP-Label.