{"id":"W4402262800","doi":"10.1109/igarss53475.2024.10641520","title":"UnderstAnding Bag of Tricks of Deep Learning-Based Semantic Segmentation in Pavement Crack Detection","year":2024,"lang":"en","type":"article","venue":"","topic":"Infrastructure Maintenance and Monitoring","field":"Engineering","cited_by":1,"is_retracted":false,"has_abstract":true,"ca_institutions":"University of Waterloo","funders":"","keywords":"Computer science; Segmentation; Artificial intelligence; Deep learning; Image segmentation; Computer vision; Natural language processing","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":[0.0001109463,0.00006498153,0.00009139642,0.0002101753,0.00001182691,0.000009888723,0.00002458064,0.00003582029,0.00003899611],"category_scores_gemma":[0.000008024674,0.00006030827,0.00003183763,0.0002366226,0.000009223045,0.00006564557,0.000004795201,0.00009683232,0.000001550605],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0002147381,"about_ca_system_score_gemma":0.00000628021,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00003136072,"about_ca_topic_score_gemma":0.00007576631,"domain_scores_codex":[0.9995244,0.00001074422,0.0001931327,0.00007242648,0.00009717528,0.0001021751],"domain_scores_gemma":[0.9998739,0.00003833591,0.00001904422,0.00004478114,0.00001251872,0.00001146305],"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":[0.000007806037,0.000006154906,0.002157342,0.0006423243,0.00002643167,0.000003945223,0.0006823898,0.790963,0.188044,0.0002281742,0.000006170297,0.01723226],"study_design_scores_gemma":[0.0002051322,0.00004803689,0.0009785583,0.0001295915,0.00001159616,5.636966e-7,0.001087978,0.5455953,0.451615,0.0002373968,0.00002975729,0.00006108503],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.2640287,0.0000863302,0.7347384,0.000004614483,0.0003732408,0.00008224222,3.038768e-7,0.00007543167,0.0006107574],"genre_scores_gemma":[0.9992803,0.00001638798,0.0006346625,0.000002161336,0.00002752343,0.000005130991,0.000002015991,0.00001269838,0.00001910302],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.7352517,"threshold_uncertainty_score":0.24593,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.01874564306517288,"score_gpt":0.231089385329483,"score_spread":0.2123437422643102,"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."}}