Towards Realistic Remote Sensing Dataset Distillation with Discriminative Prototype-guided Diffusion

TL;DR

This paper introduces a novel dataset distillation method for remote sensing images using a diffusion model guided by classifiers and prototypes, reducing data size while maintaining diversity and realism for training.

Contribution

It pioneers the application of dataset distillation in remote sensing with a diffusion model guided by discriminative prototypes and classification consistency.

Findings

Effective distillation of large datasets into compact, realistic samples

Improved diversity and discriminative quality of synthesized images

Successful application on high-resolution remote sensing benchmarks

Abstract

Recent years have witnessed the remarkable success of deep learning in remote sensing image interpretation, driven by the availability of large-scale benchmark datasets. However, this reliance on massive training data also brings two major challenges: (1) high storage and computational costs, and (2) the risk of data leakage, especially when sensitive categories are involved. To address these challenges, this study introduces the concept of dataset distillation into the field of remote sensing image interpretation for the first time. Specifically, we train a text-to-image diffusion model to condense a large-scale remote sensing dataset into a compact and representative distilled dataset. To improve the discriminative quality of the synthesized samples, we propose a classifier-driven guidance by injecting a classification consistency loss from a pre-trained model into the diffusion training process. Besides, considering the rich semantic complexity of remote sensing imagery, we further perform latent space clustering on training samples to select representative and diverse prototypes as visual style guidance, while using a visual language model to provide aggregated text descriptions. Experiments on three high-resolution remote sensing scene classification benchmarks show that the proposed method can distill realistic and diverse samples for downstream model training. Code and pre-trained models are available online (https://github.com/YonghaoXu/DPD).