OverLayBench: A Benchmark for Layout-to-Image Generation with Dense Overlaps

Jianwen Xie (Lambda) · Bingnan Li (University of California, San Diego) · Chen-Yu Wang (University of California, San Diego) · Haiyang Xu (University of California, San Diego) · Xiang Zhang (University of California, San Diego) · Ethan Armand (University of California, San Diego) · Divyansh Srivastava (University of California, San Diego) · Shan Xiaojun · Zeyuan Chen (University of California, San Diego) · Zhuowen Tu (University of California, San Diego)

amodal mask datasetchallenging scenarioscomplexity quantificationgeneration qualityhigh-quality annotationslayout-to-image generation model evaluationoverlapping bounding boxesoverlaybenchoverlayscoreperformance improvementqualitative examplesquantitative analysisrobust generationsemantic distinction

Despite steady progress in layout-to-image generation, current methods still struggle with layouts containing significant overlap between bounding boxes. We identify two primary challenges: (1) large overlapping regions and (2) overlapping instances with minimal semantic distinction. Through both qualitative examples and quantitative analysis, we demonstrate how these factors degrade generation quality. To systematically assess this issue, we introduce OverLayScore, a novel metric that quantifies the complexity of overlapping bounding boxes. Our analysis reveals that existing benchmarks are biased toward simpler cases with low OverLayScore values, limiting their effectiveness in evaluating models under more challenging conditions. To reduce this gap, we present OverLayBench, a new benchmark featuring balanced OverLayScore distributions and high-quality annotations. As an initial step toward improved performance on complex overlaps, we also propose CreatiLayout-AM, a model trained on a curated amodal mask dataset. Together, our contributions establish a foundation for more robust layout-to-image generation under realistic and challenging scenarios.