The paper "TCI-Former: Thermal Conduction-Inspired Transformer for Infrared Small Target Detection" addresses the challenge of infrared small target detection (ISTD), a critical application in national security and military areas. The authors propose a novel approach, TCI-Former, which draws inspiration from thermal conduction in thermodynamics to enhance ISTD performance. Key contributions of the paper include: 1. **Deriving the Pixel Movement Differential Equation (PMDE)**: Inspired by the thermal conduction differential equation, the authors derive the PMDE to model the evolution of feature maps during ISTD, linking spatial and temporal information. 2. **TCI-Former Architecture**: The TCI-Former is designed with a U-Net-like encoder-decoder structure, incorporating two key modules: Thermal Conduction-Inspired Attention (TCIA) and Thermal Conduction Boundary Module (TCBM). TCIA uses the finite difference method to extract target body features, while TCBM refines these features with fine boundary details. 3. **Loss Function**: A hybrid loss function combining segmentation loss, target boundary loss, and interior body loss is used to optimize the model. 4. **Experiments and Results**: The method is evaluated on the NUAA-SIRST and IRSTD-1k datasets, demonstrating superior performance in both pixel-level and object-level metrics compared to state-of-the-art methods. The paper highlights the effectiveness of the TCI-Former in achieving accurate and boundary-aware segmentation of small targets, making it a significant advancement in ISTD.The paper "TCI-Former: Thermal Conduction-Inspired Transformer for Infrared Small Target Detection" addresses the challenge of infrared small target detection (ISTD), a critical application in national security and military areas. The authors propose a novel approach, TCI-Former, which draws inspiration from thermal conduction in thermodynamics to enhance ISTD performance. Key contributions of the paper include: 1. **Deriving the Pixel Movement Differential Equation (PMDE)**: Inspired by the thermal conduction differential equation, the authors derive the PMDE to model the evolution of feature maps during ISTD, linking spatial and temporal information. 2. **TCI-Former Architecture**: The TCI-Former is designed with a U-Net-like encoder-decoder structure, incorporating two key modules: Thermal Conduction-Inspired Attention (TCIA) and Thermal Conduction Boundary Module (TCBM). TCIA uses the finite difference method to extract target body features, while TCBM refines these features with fine boundary details. 3. **Loss Function**: A hybrid loss function combining segmentation loss, target boundary loss, and interior body loss is used to optimize the model. 4. **Experiments and Results**: The method is evaluated on the NUAA-SIRST and IRSTD-1k datasets, demonstrating superior performance in both pixel-level and object-level metrics compared to state-of-the-art methods. The paper highlights the effectiveness of the TCI-Former in achieving accurate and boundary-aware segmentation of small targets, making it a significant advancement in ISTD.