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

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Wang, Yibo, Gao, Ruiyuan, Chen, Kai, Zhou, Kaiqiang, Cai, Yingjie, Hong, Lanqing, Li, Zhenguo, Jiang, Lihui, Yeung, Dit-Yan, Xu, Qiang, Zhang, Kai
Format: Preprint
Published: 2024
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866911804612935680
author Wang, Yibo
Gao, Ruiyuan
Chen, Kai
Zhou, Kaiqiang
Cai, Yingjie
Hong, Lanqing
Li, Zhenguo
Jiang, Lihui
Yeung, Dit-Yan
Xu, Qiang
Zhang, Kai
author_facet Wang, Yibo
Gao, Ruiyuan
Chen, Kai
Zhou, Kaiqiang
Cai, Yingjie
Hong, Lanqing
Li, Zhenguo
Jiang, Lihui
Yeung, Dit-Yan
Xu, Qiang
Zhang, Kai
contents 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.
format Preprint
id arxiv_https___arxiv_org_abs_2403_13304
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle DetDiffusion: Synergizing Generative and Perceptive Models for Enhanced Data Generation and Perception
Wang, Yibo
Gao, Ruiyuan
Chen, Kai
Zhou, Kaiqiang
Cai, Yingjie
Hong, Lanqing
Li, Zhenguo
Jiang, Lihui
Yeung, Dit-Yan
Xu, Qiang
Zhang, Kai
Computer Vision and Pattern Recognition
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.
title DetDiffusion: Synergizing Generative and Perceptive Models for Enhanced Data Generation and Perception
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2403.13304