RealScene-ISTD
收藏资源简介:
RealScene-ISTD是由广东工业大学信息工程学院创建的一个新型跨域红外小目标数据集,旨在促进模型在跨域数据条件下的泛化研究以及现有红外小目标检测算法在真实世界环境中的性能优化。该数据集包含了多种形状、类别、姿态和来自多个来源的传感器模态,为评估模型在跨域中的泛化能力和推动红外小目标检测算法在真实世界应用中的性能提升提供了基准。
RealScene-ISTD is a novel cross-domain infrared small target dataset developed by the School of Information Engineering, Guangdong University of Technology. It aims to facilitate research on model generalization under cross-domain data conditions and optimize the performance of existing infrared small target detection algorithms in real-world environments. This dataset encompasses diverse shapes, categories, poses and sensor modalities from multiple sources, providing a benchmark for evaluating the cross-domain generalization capability of models and advancing the performance enhancement of infrared small target detection algorithms in real-world applications.

- 1Rethinking Generalizable Infrared Small Target Detection: A Real-scene Benchmark and Cross-view Representation Learning广东工业大学信息工程学院 · 2025年



