IEEE Transactions on Geoscience and Remote Sensing, 2024
Fig. 1: We compared the effects of the traditional Mosaic [1] with our method. To emphasize the diversity of samples generated by our method, we fix the top-right image of Mosaic when combining rest four images. The samples generated by the Mosaic exhibit a fragmented quality and fail to resemble a complete image. In contrast, Diff-Mosaic has a uniform distribution and coordinated grayscale. Especially in terms of the infrared small targets marked with red circles, the results of Diff-Mosaic can better enhance the diversity and realism.
Overview
Introduces a diffusion-prior-based data augmentation method with pixel harmonization and diffusion enhancement to generate more realistic and diverse infrared small-target samples.
