Detect object by train on text-to-image for foreground and background with Text2Image-for-Detection
Detect object by train on text-to-image for foreground and background with Text2Image-for-Detection
Beyond Generation: Harnessing Text to Image Models for Object Detection and Segmentation
arXiv paper abstract https://arxiv.org/abs/2309.05956
arXiv PDF paper https://arxiv.org/pdf/2309.05956.pdf
... propose a new paradigm to automatically generate training data with accurate labels at scale using the text-to-image synthesis frameworks (e.g., DALL-E, Stable Diffusion, etc.).
The proposed approach decouples training data generation into foreground object generation, and contextually coherent background generation.
To generate foreground objects, ... employ ... object class name as input prompts ... into a text-to-image ... producing various foreground images ... against ... backgrounds. A foreground-background segmentation ... generate foreground object masks.
To generate context images, ... begin by creating language descriptions of the context ... by ... image captioning ... images representing ... context ... descriptions are then transformed into a diverse array of context images via a text-to-image synthesis framework.
... composite these with the foreground object masks produced in the initial step, utilizing a cut-and-paste method, to formulate the training data.
... detectors trained solely on synthetic data produced by ... method achieve performance comparable to those trained on real data ...
Please like and share this post if you enjoyed it using the buttons at the bottom!
Stay up to date. Subscribe to my posts https://morrislee1234.wixsite.com/website/contact
Web site with my other posts by category https://morrislee1234.wixsite.com/website
Comments