IRSTFormer: A Hierarchical Vision Transformer for Infrared Small Target Detection

Chen, Gao; Tan, Sirui

doi:10.3390/rs14143258

Cited by 27 publications

(21 citation statements)

References 46 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Due to variations in scanning parameters, MRI data collected from different patients may exhibit variations in size, intensity, and orientation. The dataset can be expanded and diversified by implementing data augmentation strategies, such as random rotation, flipping, intensity scaling, and intensity shifting (Chen et al 2018). Table 7 shows the segmentation performance when using different combinations of data enhancement strategies.…”

Section: Comparison Of Data Augmentation Strategiesmentioning

confidence: 99%

Sliding transformer with uncertainty estimation for vestibular schwannoma automatic segmentation

Liu,

Li,

et al. 2024

Phys. Med. Biol.

View full text Add to dashboard Cite

Objective. Automated segmentation of vestibular schwannoma (VS) using magnetic resonance imaging (MRI) can enhance clinical efficiency. Though many advanced methods exist for automated VS segmentation, the accuracy is hindered by ambivalent tumor borders and cystic regions in some patients. In addition, these methods provide results that do not indicate segmentation uncertainty, making their translation into clinical workflows difficult due to potential errors. Providing a definitive segmentation result along with segmentation uncertainty or self-confidence is crucial for the conversion of automated segmentation programs to clinical aid diagnostic tools. Approach. To address these issues, we propose a U-shaped cascade transformer structure with a sliding window that utilizes multiple sliding samples, a segmentation head, and an uncertainty head to obtain both the segmentation mask and uncertainty map. We collected multimodal MRI data from 60 clinical patients with VS from Xuanwu Hospital. Each patient case includes T1-weighted images, contrast-enhanced T1-weighted images, T2-weighted images, and a tumor mask. The images exhibit an in-plane resolution ranging from 0.70×0.70 to 0.76×0.76 mm, an in-plane matrix spanning from 216×256 to 284×256, a slice thickness varying between 0.50 and 0.80 mm, and a range of slice numbers from 72 to 120. Main results. Extensive experimental results show that our method achieves comparable or higher results than previous state-of-the-art brain tumor segmentation methods. On our collected multimodal MRI dataset of clinical VS, our method achieved the Dice similarity coefficient (DSC) of 96.08%±1.30. On a publicly available VS dataset, our method achieved the mean DSC of 94.23%±2.53. Significance. The method efficiently solves the VS segmentation task while providing an uncertainty map of the segmentation results, which helps clinical experts review the segmentation results more efficiently and helps to transform the automated segmentation program into a clinical aid diagnostic tool.

show abstract

Section: Comparison Of Data Augmentation Strategiesmentioning

confidence: 99%

Sliding transformer with uncertainty estimation for vestibular schwannoma automatic segmentation

Liu,

Li,

et al. 2024

Phys. Med. Biol.

View full text Add to dashboard Cite

show abstract

“…Infrared Small Target Detection Networks However, only local features are insufficient to detect all infrared targets because the low contrast background makes many small targets unclear to find. Therefore, researchers turn to hybrid methods (Chen, Wang, and Tan 2022;Wang et al 2022a;Zhang et al 2022a) by combining ViT with CNN to complement local features with global dependencies. For example, Chen et al novelly built a ViT-CNN structure based on fluid dynamics for shape-aware ISTD.…”

Section: Related Workmentioning

confidence: 99%

TCI-Former: Thermal Conduction-Inspired Transformer for Infrared Small Target Detection

Chen,

Tan,

Chu

et al. 2024

AAAI

View full text Add to dashboard Cite

Infrared small target detection (ISTD) is critical to national security and has been extensively applied in military areas. ISTD aims to segment small target pixels from background. Most ISTD networks focus on designing feature extraction blocks or feature fusion modules, but rarely describe the ISTD process from the feature map evolution perspective. In the ISTD process, the network attention gradually shifts towards target areas. We abstract this process as the directional movement of feature map pixels to target areas through convolution, pooling and interactions with surrounding pixels, which can be analogous to the movement of thermal particles constrained by surrounding variables and particles. In light of this analogy, we propose Thermal Conduction-Inspired Transformer (TCI-Former) based on the theoretical principles of thermal conduction. According to thermal conduction differential equation in heat dynamics, we derive the pixel movement differential equation (PMDE) in the image domain and further develop two modules: Thermal Conduction-Inspired Attention (TCIA) and Thermal Conduction Boundary Module (TCBM). TCIA incorporates finite difference method with PMDE to reach a numerical approximation so that target body features can be extracted. To further remove errors in boundary areas, TCBM is designed and supervised by boundary masks to refine target body features with fine boundary details. Experiments on IRSTD-1k and NUAA-SIRST demonstrate the superiority of our method.

show abstract

“…With the development of deep learning, some data-driven methods and infrared small target datasets [7,[19][20][21][22] have emerged in recent years. Considering the weak and small characteristics of infrared small targets, infrared small target detection is usually modeled as a semantic segmentation problem.…”

Section: Introductionmentioning

confidence: 99%

CCRANet: A Two-Stage Local Attention Network for Single-Frame Low-Resolution Infrared Small Target Detection

Wang,

Xiao,

Dou

et al. 2023

Remote Sensing

View full text Add to dashboard Cite

Infrared small target detection technology is widely used in infrared search and tracking, infrared precision guidance, low and slow small aircraft detection, and other projects. Its detection ability is very important in terms of finding unknown targets as early as possible, warning in time, and allowing for enough response time for the security system. This paper combines the target characteristics of low-resolution infrared small target images and studies the infrared small target detection method under a complex background based on the attention mechanism. The main contents of this paper are as follows: (1) by sorting through and expanding the existing datasets, we construct a single-frame low-resolution infrared small target (SLR-IRST) dataset and evaluate the existing datasets on three aspects—target number, target category, and target size; (2) to improve the pixel-level metrics of low-resolution infrared small target detection, we propose a small target detection network with two stages and a corresponding method. Regarding the SLR-IRST dataset, the proposed method is superior to the existing methods in terms of pixel-level metrics and target-level metrics and has certain advantages in model processing speed.

show abstract

IRSTFormer: A Hierarchical Vision Transformer for Infrared Small Target Detection

Cited by 27 publications

References 46 publications

Sliding transformer with uncertainty estimation for vestibular schwannoma automatic segmentation

Sliding transformer with uncertainty estimation for vestibular schwannoma automatic segmentation

TCI-Former: Thermal Conduction-Inspired Transformer for Infrared Small Target Detection

CCRANet: A Two-Stage Local Attention Network for Single-Frame Low-Resolution Infrared Small Target Detection

Contact Info

Product

Resources

About