{"ID":5935659,"CreatedAt":"2026-07-07T01:22:02.77346169Z","UpdatedAt":"2026-07-07T02:10:06.972658124Z","DeletedAt":null,"paper_url":"https://arxiv.org/abs/2607.03488","arxiv_id":"2607.03488","title":"Learning to Generate Multiple Objects from Dense and Occluded Layouts","abstract":"Text-to-image diffusion models fail to generate correct object counts in dense scenes, where overlapping instances collapse into indistinguishable structures despite appearing visually plausible. We identify this as instance ownership collapse: tokens from overlapping objects interact freely through attention, while heavily occluded instances receive weak supervision due to their small visible areas. We address this through layout-aware attention biases that softly bias token interactions toward region-consistent grouping and suppress cross-instance leakage, paired with an amodal-balanced loss that amplifies gradients for occluded objects based on their occlusion level. To enable systematic evaluation, we introduce OverlapDepth-45K, a benchmark of densely overlapping scenes with amodal supervision. Our approach substantially improves count accuracy and prevents instance merging while preserving image quality. Project page: https://bachngoh.github.io/AIBL","short_abstract":"Text-to-image diffusion models fail to generate correct object counts in dense scenes, where overlapping instances collapse into indistinguishable structures despite appearing visually plausible. We identify this as instance ownership collapse: tokens from overlapping objects interact freely through attention, while he...","url_abs":"https://arxiv.org/abs/2607.03488","url_pdf":"https://arxiv.org/pdf/2607.03488v1","authors":"[\"Bach-Hoang Ngo\",\"Si-Tri Ngo\",\"Hieu Le\",\"Trung-Nghia Le\"]","published":"2026-07-03T16:56:01Z","proceeding":"cs.CV","tasks":"[\"cs.CV\"]","methods":"[\"Diffusion Model\"]","has_code":false}