Dataset Groups Activity Stream SceneComposer: Any-Level Semantic Image Synthesis A new framework for conditional image synthesis from semantic layouts of any precision levels, ranging from pure text to a 2D semantic canvas with precise shapes. BibTex: @dataset{Yu_Zeng1_and_Zhe_Lin2_and_Jianming_Zhang2_and_Qing_Liu2_and_John_Collomosse2_and_Jason_Kuen2_and_Vishal_M_Patel1_2024, abstract = {A new framework for conditional image synthesis from semantic layouts of any precision levels, ranging from pure text to a 2D semantic canvas with precise shapes.}, author = {Yu Zeng1 and Zhe Lin2 and Jianming Zhang2 and Qing Liu2 and John Collomosse2 and Jason Kuen2 and Vishal M. Patel1}, doi = {10.57702/7kv69m9c}, institution = {No Organization}, keyword = {'layout control', 'precision levels', 'semantic image synthesis'}, month = {dec}, publisher = {TIB}, title = {SceneComposer: Any-Level Semantic Image Synthesis}, url = {https://service.tib.eu/ldmservice/dataset/scenecomposer--any-level-semantic-image-synthesis}, year = {2024} }