{"id":"W4400979646","doi":"10.1364/boe.524603","title":"Pretraining of 3D image segmentation models for retinal OCT using denoising-based self-supervised learning","year":2024,"lang":"en","type":"article","venue":"Biomedical Optics Express","topic":"Retinal Imaging and Analysis","field":"Medicine","cited_by":4,"is_retracted":false,"has_abstract":true,"ca_institutions":"Association of Canadian Universities for Northern Studies","funders":"Heidelberg Engineering; Christian Doppler Forschungsgesellschaft; Österreichische Nationalstiftung für Forschung, Technologie und Entwicklung","keywords":"Segmentation; Computer science; Artificial intelligence; Image segmentation; Noise reduction; Pattern recognition (psychology); Scale-space segmentation; Automation; Deep learning; Supervised learning; Computer vision; Machine learning; Segmentation-based object categorization; Training set; Set (abstract data type); Artificial neural network","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":[0.0006015435,0.0001768431,0.0003691222,0.0002971168,0.00009627657,0.00008280404,0.0001008427,0.0001189178,0.00003633393],"category_scores_gemma":[0.0002290053,0.000153231,0.0002047165,0.0003938186,0.0001757111,0.0001437473,0.00003249757,0.0002592171,0.0000024285],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00008363633,"about_ca_system_score_gemma":0.0002189636,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00001689115,"about_ca_topic_score_gemma":7.886882e-8,"domain_scores_codex":[0.9982509,0.0000656526,0.0004681147,0.0003523859,0.0005626843,0.0003002579],"domain_scores_gemma":[0.9990545,0.0002787201,0.00009924878,0.0001752218,0.0002025114,0.000189823],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"bench_or_experimental","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.0002817834,0.0003294882,0.0003993749,0.003398298,0.0004052971,0.0001281263,0.00224674,0.006198535,0.970537,0.0001710037,0.0002548033,0.01564957],"study_design_scores_gemma":[0.0009786965,0.0003186821,0.00001008059,0.001246627,0.0007013806,0.00002353943,0.0004143082,0.9771891,0.01859241,0.0000700543,0.0003038425,0.0001512415],"study_design_candidate":"bench_or_experimental","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.2528501,0.000416623,0.7454987,0.0003563687,0.0001229237,0.00024928,0.00001918844,0.0001735451,0.0003132864],"genre_scores_gemma":[0.5845855,0.00001811695,0.4149068,0.00005362214,0.0001964528,0.00001263598,0.0001121341,0.00003626309,0.00007841478],"genre_candidate":"methods","genre_consensus":null,"teacher_disagreement_score":0.9709906,"threshold_uncertainty_score":0.6248577,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.03374584944137498,"score_gpt":0.3285330322601953,"score_spread":0.2947871828188203,"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."}}