{"id":"W3098535516","doi":"10.1016/j.media.2021.102038","title":"SoftSeg: Advantages of soft versus binary training for image segmentation","year":2021,"lang":"en","type":"preprint","venue":"Medical Image Analysis","topic":"Domain Adaptation and Few-Shot Learning","field":"Computer Science","cited_by":3,"is_retracted":false,"has_abstract":false,"ca_institutions":"Polytechnique Montréal; Mila - Quebec Artificial Intelligence Institute","funders":"Fonds de recherche du Québec – Nature et technologies; Fonds de Recherche du Québec - Santé; Canadian Institutes of Health Research; Natural Sciences and Engineering Research Council of Canada; Institut de Valorisation des Données; Canada First Research Excellence Fund; Canada Research Chairs; Canada Foundation for Innovation; Nvidia","keywords":"Artificial intelligence; Segmentation; Computer science; Pattern recognition (psychology); Voxel; Preprocessor; Binary classification; Image segmentation; Binary number; Ground truth; Pixel; Mathematics; Support vector machine","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":["metaepi_narrow"],"consensus_categories":[],"category_scores_codex":[0.001286059,0.0002912284,0.0008347472,0.000594245,0.0001361387,0.0003195308,0.001099892,0.0002577429,0.0006593671],"category_scores_gemma":[0.001860161,0.0002931216,0.000851025,0.001161131,0.0001887189,0.0004677573,0.0008307184,0.0005455249,0.00001021693],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00007326889,"about_ca_system_score_gemma":0.0005068132,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00004962834,"about_ca_topic_score_gemma":0.00003275588,"domain_scores_codex":[0.9965634,0.0002652867,0.0007650335,0.0008549187,0.00117868,0.0003726394],"domain_scores_gemma":[0.9971548,0.0009009348,0.0005317578,0.0007407274,0.0003948188,0.0002769455],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.0002998104,0.0008291642,0.0006730533,0.001399931,0.01122951,0.00104997,0.02862507,0.01813156,0.02088812,0.001223917,0.002227303,0.9134226],"study_design_scores_gemma":[0.001640191,0.0001152338,0.0008775242,0.0001471707,0.001583918,0.00000443516,0.00432068,0.9883269,0.001559514,0.0003073503,0.0006302178,0.000486874],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.008593353,0.0003877187,0.9882125,0.001333559,0.0005137476,0.0002026702,0.00002043396,0.0001311402,0.0006049356],"genre_scores_gemma":[0.2760752,0.000215492,0.7215362,0.0004789816,0.0002083656,0.0001223576,0.0009526757,0.00003584693,0.0003749159],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.9701954,"threshold_uncertainty_score":0.9999521,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.04295141484576424,"score_gpt":0.3517031474118046,"score_spread":0.3087517325660404,"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."}}