DetDiffusion: Synergizing Generative and Perceptive Models for Enhanced Data Generation and Perception

TL;DR

DetDiffusion innovatively combines generative and perceptive models using perception-aware loss and attribute-based augmentation to improve data generation quality and perception tasks, achieving state-of-the-art results in layout-guided generation and detection.

Contribution

It introduces a novel framework that harmonizes generative and perceptive models, utilizing perception-aware loss and attribute-based augmentation for enhanced data synthesis and perception performance.

Findings

Achieves state-of-the-art in layout-guided generation

Significantly improves downstream detection performance

Enhances image quality and controllability through perception-aware techniques

Abstract

Current perceptive models heavily depend on resource-intensive datasets, prompting the need for innovative solutions. Leveraging recent advances in diffusion models, synthetic data, by constructing image inputs from various annotations, proves beneficial for downstream tasks. While prior methods have separately addressed generative and perceptive models, DetDiffusion, for the first time, harmonizes both, tackling the challenges in generating effective data for perceptive models. To enhance image generation with perceptive models, we introduce perception-aware loss (P.A. loss) through segmentation, improving both quality and controllability. To boost the performance of specific perceptive models, our method customizes data augmentation by extracting and utilizing perception-aware attribute (P.A. Attr) during generation. Experimental results from the object detection task highlight DetDiffusion's superior performance, establishing a new state-of-the-art in layout-guided generation. Furthermore, image syntheses from DetDiffusion can effectively augment training data, significantly enhancing downstream detection performance.