{"id":"W4353055872","doi":"10.1016/j.cmpb.2023.107505","title":"Dual center validation of deep learning for automated multi-label segmentation of thoracic anatomy in bedside chest radiographs","year":2023,"lang":"en","type":"article","venue":"Computer Methods and Programs in Biomedicine","topic":"COVID-19 diagnosis using AI","field":"Medicine","cited_by":5,"is_retracted":false,"has_abstract":false,"ca_institutions":"Artificial Intelligence in Medicine (Canada)","funders":"Berlin Institute of Health; Charité – Universitätsmedizin Berlin","keywords":"Segmentation; Medicine; Jaccard index; Deep learning; Artificial intelligence; Convolutional neural network; Sørensen–Dice coefficient; Computer science; Radiography; Workflow; Radiology; Pattern recognition (psychology); Image segmentation","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":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.002277416,0.000158337,0.0005304522,0.001178993,0.00002783004,0.00001120864,0.00006101578,0.0001122965,0.000002465394],"category_scores_gemma":[0.000175835,0.0001387044,0.00005985673,0.001997486,0.0001650492,0.0000646005,0.0000612537,0.0001455296,2.837668e-7],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00005945616,"about_ca_system_score_gemma":0.00002616297,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0001366023,"about_ca_topic_score_gemma":0.00001126756,"domain_scores_codex":[0.9983032,0.0003150458,0.0006299344,0.0003193404,0.0001878596,0.0002445684],"domain_scores_gemma":[0.9987723,0.0006489917,0.0002250355,0.0001512443,0.000127203,0.00007524258],"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":[0.0001235548,0.0004154318,0.1884465,0.0012254,0.0000529037,0.00001312157,0.001794861,0.00009053775,0.01224927,0.000005388424,0.00003653301,0.7955465],"study_design_scores_gemma":[0.007276995,0.00145312,0.1907213,0.001556066,0.0001114621,0.00001732775,0.0003704718,0.7832043,0.01379575,0.00006967664,0.001259753,0.0001637855],"study_design_candidate":"design_other","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"methods","genre_scores_codex":[0.7967594,0.0005022282,0.1995541,0.001519667,0.0002376663,0.001233965,0.000003454327,0.0001884146,0.00000106175],"genre_scores_gemma":[0.2783934,0.0001931936,0.7206691,0.0002010195,0.00006346455,0.00009543197,0.0003570394,0.00002336972,0.000003957257],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.7953827,"threshold_uncertainty_score":0.5656203,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.09771304373009508,"score_gpt":0.4705496976380069,"score_spread":0.3728366539079118,"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."}}