{"id":"W3134978274","doi":"10.3389/fmed.2021.629134","title":"The Effectiveness of Image Augmentation in Deep Learning Networks for Detecting COVID-19: A Geometric Transformation Perspective","year":2021,"lang":"en","type":"article","venue":"Frontiers in Medicine","topic":"COVID-19 diagnosis using AI","field":"Medicine","cited_by":113,"is_retracted":false,"has_abstract":true,"ca_institutions":"Vancouver General Hospital; University of Manitoba; Simon Fraser University; University of British Columbia","funders":"Natural Sciences and Engineering Research Council of Canada; Canada Research Chairs","keywords":"Overfitting; Deep learning; Artificial intelligence; Computer science; Coronavirus disease 2019 (COVID-19); Transformation (genetics); Machine learning; Training set; Perspective (graphical); Artificial neural network; Image (mathematics); Geometric transformation; Early stopping; Pattern recognition (psychology); Medicine","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":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.003065842,0.001953753,0.0007996773,0.0007542532,0.0003792495,0.001122407,0.001195425,0.001429049,0.0009483909],"category_scores_gemma":[0.01129656,0.0004663937,0.0009011431,0.0005255463,0.0009815716,0.002301689,0.001342946,0.002243963,0.0005061523],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0006644898,"about_ca_system_score_gemma":0.001046921,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.004776541,"about_ca_topic_score_gemma":0.003833634,"domain_scores_codex":[0.9987348,0.0004561743,0.00007844034,0.0002836482,0.0002896645,0.0001573157],"domain_scores_gemma":[0.9961327,0.002298755,0.0003386368,0.0005454076,0.0005334057,0.0001509782],"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":[0.001154604,0.0004793663,0.02052304,0.0002961304,0.0003164317,0.0003361518,0.0001800532,0.5172685,0.02812395,0.003809576,0.005990433,0.4215218],"study_design_scores_gemma":[0.00002035853,0.0002523227,0.001791808,0.00003365114,0.00005228973,0.0001044869,0.00002229969,0.9840569,0.01087339,0.001783906,0.0009904029,0.00001802606],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.6119768,0.007517901,0.3599336,0.003425469,0.0005773262,0.0002199802,0.0006335804,0.006176064,0.009539311],"genre_scores_gemma":[0.8957273,0.001125736,0.09945652,0.0006255095,0.0001720863,0.00008581309,0.0008309283,0.0001281984,0.001847784],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.004776541,"threshold_uncertainty_score":0.01621389,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.01561574484348434,"score_gpt":0.3347402999070178,"score_spread":0.3191245550635335,"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."}}