{"id":"W4385190584","doi":"10.1016/j.neuroimage.2023.120289","title":"ROOD-MRI: Benchmarking the robustness of deep learning segmentation models to out-of-distribution and corrupted data in MRI","year":2023,"lang":"en","type":"article","venue":"NeuroImage","topic":"Advanced Neural Network Applications","field":"Computer Science","cited_by":32,"is_retracted":false,"has_abstract":true,"ca_institutions":"Baycrest Hospital; Heart and Stroke Foundation; University of Toronto; Western University; Sunnybrook Health Science Centre","funders":"Natural Sciences and Engineering Research Council of Canada; Government of Ontario; Canada Research Chairs; Compute Canada; Temerty Family Foundation; Ontario Brain Institute; University of Ottawa","keywords":"Benchmarking; Robustness (evolution); Segmentation; Computer science; Artificial intelligence; Benchmark (surveying); Deep learning; Pattern recognition (psychology); Machine learning; Data mining; Cartography","routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false,"invisible_to_affiliation_only":false},"retraction":null,"screen":null,"direct_labels":[],"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0002820282,0.00008077961,0.0001112751,0.00006715633,0.0000994234,0.00003361411,0.0006562591,0.0000216327,0.000001112577],"category_scores_gemma":[0.00005064325,0.0000730574,0.00001284704,0.0009874136,0.00004209148,0.0004983658,0.0007101202,0.000140281,0.000002548062],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00001430839,"about_ca_system_score_gemma":0.00001071149,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00001087833,"about_ca_topic_score_gemma":0.00003921191,"domain_scores_codex":[0.9989558,0.00009504714,0.000234066,0.0003639772,0.0001869082,0.0001641586],"domain_scores_gemma":[0.9989176,0.000254883,0.0001272527,0.0006178441,0.00004931974,0.00003305392],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.000006104837,0.00001596011,0.0007535712,0.00001352853,0.000002077294,0.000002719743,0.0005024621,0.9459393,0.00732059,0.0009217262,0.000154145,0.04436779],"study_design_scores_gemma":[0.0001221733,0.00003084016,0.0135156,0.00001433186,0.000003617728,0.000001843782,0.0000564646,0.9849981,0.0006932476,0.0003680102,0.0001339792,0.00006181457],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.1137518,0.00002123404,0.8851703,0.0005359743,0.000112103,0.0002873091,0.00001543933,0.00006451825,0.0000413467],"genre_scores_gemma":[0.9814603,0.00006827633,0.01821768,0.00004574166,0.0000275745,0.00002377829,0.0001346585,0.000008490747,0.00001346619],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.8677085,"threshold_uncertainty_score":0.2979194,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.06557973590540603,"score_gpt":0.3127212765602763,"score_spread":0.2471415406548702,"validation_status":"score_only:v0-immature-baseline","note":"Baseline scores from an immature model (maturity gate not passed). Scores rank; they never assert a category."}}