{"id":"W3153741272","doi":"10.1371/journal.pone.0255809","title":"Colonoscopy polyp detection and classification: Dataset creation and comparative evaluations","year":2021,"lang":"en","type":"preprint","venue":"PLoS ONE","topic":"Colorectal Cancer Screening and Detection","field":"Medicine","cited_by":6,"is_retracted":false,"has_abstract":true,"ca_institutions":"Toronto Metropolitan University","funders":"National Cancer Institute; National Institutes of Health","keywords":"Benchmark (surveying); Colonoscopy; Artificial intelligence; Computer science; Ground truth; Colorectal cancer; Deep learning; Machine learning; Pattern recognition (psychology); Cancer; Medicine; Internal medicine","routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false,"invisible_to_affiliation_only":false},"retraction":null,"screen":null,"direct_labels":[],"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.004052011,0.002265771,0.001279249,0.004253705,0.001007211,0.001503195,0.002563816,0.002309574,0.002980253],"category_scores_gemma":[0.00895366,0.000346903,0.001854359,0.003035835,0.001047392,0.001219847,0.002248318,0.001313093,0.002890105],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001684694,"about_ca_system_score_gemma":0.001767515,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.01958391,"about_ca_topic_score_gemma":0.02836007,"domain_scores_codex":[0.9957208,0.0008778447,0.0005781171,0.0009493456,0.001474896,0.0003991461],"domain_scores_gemma":[0.9950369,0.001238509,0.0004218274,0.001341677,0.00140445,0.0005565993],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"observational","study_design_scores_codex":[0.006742821,0.007042284,0.06734288,0.007733138,0.002045232,0.002033612,0.0004250064,0.02797046,0.03055845,0.00229243,0.412632,0.4331817],"study_design_scores_gemma":[0.004095295,0.00616006,0.3470482,0.001586664,0.001999805,0.01209347,0.002373407,0.2038328,0.06740791,0.00477556,0.3477061,0.000920782],"study_design_candidate":"observational","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.51466,0.01143067,0.03184279,0.003332412,0.002372253,0.004574971,0.4026419,0.01473615,0.01440883],"genre_scores_gemma":[0.1695563,0.002156066,0.04630343,0.0007165783,0.0003461087,0.001355592,0.774443,0.0004967026,0.00462632],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.01958391,"threshold_uncertainty_score":0.03893983,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.2189301140687204,"score_gpt":0.3750491237013577,"score_spread":0.1561190096326373,"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."}}