iSmallNet: Densely Nested Network with Label Decoupling for Infrared Small Target Detection

TL;DR

iSmallNet is a novel densely nested network with label decoupling designed to improve infrared small target detection by effectively leveraging shape information and multi-scale context, outperforming existing methods.

Contribution

The paper introduces iSmallNet, a multi-stream densely nested network with label decoupling and new modules for enhanced infrared small target detection.

Findings

Outperforms 11 state-of-the-art detectors on NUAA-SIRST and NUDT-SIRST datasets.

Effectively maintains small targets in deep layers through multi-scale nested interaction.

Utilizes label decoupling to address boundary unbalance in small object detection.

Abstract

Small targets are often submerged in cluttered backgrounds of infrared images. Conventional detectors tend to generate false alarms, while CNN-based detectors lose small targets in deep layers. To this end, we propose iSmallNet, a multi-stream densely nested network with label decoupling for infrared small object detection. On the one hand, to fully exploit the shape information of small targets, we decouple the original labeled ground-truth (GT) map into an interior map and a boundary one. The GT map, in collaboration with the two additional maps, tackles the unbalanced distribution of small object boundaries. On the other hand, two key modules are delicately designed and incorporated into the proposed network to boost the overall performance. First, to maintain small targets in deep layers, we develop a multi-scale nested interaction module to explore a wide range of context information. Second, we develop an interior-boundary fusion module to integrate multi-granularity information. Experiments on NUAA-SIRST and NUDT-SIRST clearly show the superiority of iSmallNet over 11 state-of-the-art detectors.