{"id":"W2982143932","doi":"10.1109/iccv.2019.00033","title":"What Else Can Fool Deep Learning? Addressing Color Constancy Errors on Deep Neural Network Performance","year":2019,"lang":"en","type":"article","venue":"","topic":"Image Enhancement Techniques","field":"Computer Science","cited_by":121,"is_retracted":false,"has_abstract":true,"ca_institutions":"York University","funders":"","keywords":"Computer science; Color constancy; Artificial intelligence; Robustness (evolution); Deep neural networks; Deep learning; Artificial neural network; Color balance; Segmentation; Computer vision; Image segmentation; Pattern recognition (psychology); Image (mathematics); Image processing; Color image","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":[0.003291779,0.001613629,0.0005653926,0.0004648139,0.0005653506,0.001501089,0.001542395,0.001965356,0.00368033],"category_scores_gemma":[0.02161652,0.0003717917,0.0003695262,0.0005458728,0.001733794,0.005383105,0.00166197,0.003445934,0.001114204],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0008614237,"about_ca_system_score_gemma":0.0008487623,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.003566586,"about_ca_topic_score_gemma":0.005688186,"domain_scores_codex":[0.9990513,0.0002977375,0.00004726484,0.0002131341,0.0002613863,0.0001291644],"domain_scores_gemma":[0.9958194,0.002073121,0.0003849643,0.00101878,0.0005327492,0.000170845],"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":[0.0008530644,0.0002725546,0.01202811,0.0006271988,0.0003698134,0.0005507416,0.0005820251,0.1414047,0.07051119,0.03067082,0.03064888,0.7114809],"study_design_scores_gemma":[0.0001241728,0.0008280992,0.005783391,0.0005528187,0.0002034523,0.0007474161,0.0003661531,0.7485636,0.1125116,0.1017002,0.02850165,0.000117312],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.2994069,0.02045555,0.5957702,0.04906049,0.001834136,0.0001533529,0.0005861178,0.01291321,0.01982005],"genre_scores_gemma":[0.8765632,0.003560014,0.1076337,0.004365528,0.0005043611,0.00007281426,0.0002372372,0.0007845199,0.006278647],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.00368033,"threshold_uncertainty_score":0.01740885,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.01532608444085527,"score_gpt":0.2524683634747277,"score_spread":0.2371422790338725,"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."}}