MSCA-Net: Multi-scale context aggregation network for infrared small target detection

Abstract

In complex environments, detecting tiny infrared targets has always been challenging because of the low contrast and high noise levels inherent in infrared images. These factors often lead to the loss of crucial details during feature extraction. Moreover, existing detection methods have limitations in adequately integrating global and local information, which constrains the efficiency and accuracy of infrared small target detection. To address these challenges, we propose the MSCA-Net network architecture, which enhances feature preservation and global-local fusion through three synergistic modules: the Multi-Scale Enhanced Dilated Attention Module (MSEDA), the Positional Convolutional Block Attention Module (PCBAM), and the Channel Aggregation Module (CAB). Specifically, the MSEDA captures contextual information across different receptive fields, preserving fine-grained features. The PCBAM improves spatial understanding by modeling position-aware dependencies, while the CAB adaptively aggregates multi-level features across channels to highlight key information. Through the synergistic effect of these modules, MSCA-Net effectively retains key discriminative features and achieves robust detection performance in complex infrared scenes. The experimental results demonstrate that MSCA-Net achieves strong small target detection performance in complex backgrounds. Specifically, it attains mIoU scores of 78.43 %, 94.56 %, and 67.08 % on the NUAA-SIRST, NUDT-SIRST, and IRTSD-1K datasets, respectively, underscoring its effectiveness and suggesting potential applicability in real-world scenarios.