HCF-Net: A Hierarchical Context Fusion Network for Infrared Small Object Detection Infrared small object detection is a critical computer vision task involving the recognition and localization of tiny objects in infrared images, which often contain only a few pixels. However, due to the small size of the objects and the complex backgrounds in infrared images, this task is challenging. In this paper, we propose a deep learning method, HCF-Net, that significantly improves infrared small object detection performance through multiple practical modules. These modules include the parallelized patch-aware attention (PPA) module, dimension-aware selective integration (DASI) module, and multi-dilated channel refiner (MDCR) module. The PPA module uses a multi-branch feature extraction strategy to capture feature information at different scales and levels. The DASI module enables adaptive channel selection and fusion. The MDCR module captures spatial features of different receptive field ranges through multiple depth-separable convolutional layers. Extensive experimental results on the SIRST infrared single-frame image dataset show that the proposed HCF-Net performs well, surpassing other traditional and deep learning models. Code is available at https://github.com/zhengshuchen/HCFNet. The paper introduces HCF-Net, a deep learning model for infrared small object detection. It addresses the challenges of small object loss and low background distinctiveness in infrared small object detection. The model incorporates three key modules: PPA, DASI, and MDCR. PPA employs hierarchical feature fusion and attention mechanisms to maintain and enhance representations of small objects. DASI enhances the skip connection in U-Net, focusing on the adaptive selection and delicate fusion of high and low-dimensional features. MDCR reinforces multi-scale feature extraction and channel information representation, capturing features across various receptive field ranges. The organic combination of these modules enables the model to address the challenges of small object detection more effectively, improving detection performance and robustness. The contributions of this paper include modeling infrared small object detection as a semantic segmentation problem and proposing HCF-Net, a layer-wise context fusion network that can be trained from scratch. Three practical modules have been proposed: PPA, DASI, and MDCR. These modules effectively alleviate the issues of small object loss and low background distinctiveness in infrared small object detection. The proposed HCF-Net has been evaluated on the publicly available single-frame infrared image dataset SIRST and demonstrated a significant advantage over several state-of-the-art detection methods.HCF-Net: A Hierarchical Context Fusion Network for Infrared Small Object Detection Infrared small object detection is a critical computer vision task involving the recognition and localization of tiny objects in infrared images, which often contain only a few pixels. However, due to the small size of the objects and the complex backgrounds in infrared images, this task is challenging. In this paper, we propose a deep learning method, HCF-Net, that significantly improves infrared small object detection performance through multiple practical modules. These modules include the parallelized patch-aware attention (PPA) module, dimension-aware selective integration (DASI) module, and multi-dilated channel refiner (MDCR) module. The PPA module uses a multi-branch feature extraction strategy to capture feature information at different scales and levels. The DASI module enables adaptive channel selection and fusion. The MDCR module captures spatial features of different receptive field ranges through multiple depth-separable convolutional layers. Extensive experimental results on the SIRST infrared single-frame image dataset show that the proposed HCF-Net performs well, surpassing other traditional and deep learning models. Code is available at https://github.com/zhengshuchen/HCFNet. The paper introduces HCF-Net, a deep learning model for infrared small object detection. It addresses the challenges of small object loss and low background distinctiveness in infrared small object detection. The model incorporates three key modules: PPA, DASI, and MDCR. PPA employs hierarchical feature fusion and attention mechanisms to maintain and enhance representations of small objects. DASI enhances the skip connection in U-Net, focusing on the adaptive selection and delicate fusion of high and low-dimensional features. MDCR reinforces multi-scale feature extraction and channel information representation, capturing features across various receptive field ranges. The organic combination of these modules enables the model to address the challenges of small object detection more effectively, improving detection performance and robustness. The contributions of this paper include modeling infrared small object detection as a semantic segmentation problem and proposing HCF-Net, a layer-wise context fusion network that can be trained from scratch. Three practical modules have been proposed: PPA, DASI, and MDCR. These modules effectively alleviate the issues of small object loss and low background distinctiveness in infrared small object detection. The proposed HCF-Net has been evaluated on the publicly available single-frame infrared image dataset SIRST and demonstrated a significant advantage over several state-of-the-art detection methods.