{"id":"W2543194656","doi":"10.1109/have.2004.1391905","title":"Multimodal talker localization in video conferencing environments","year":2005,"lang":"en","type":"article","venue":"","topic":"Speech and Audio Processing","field":"Computer Science","cited_by":2,"is_retracted":false,"has_abstract":true,"ca_institutions":"Carleton University","funders":"Fonds Wetenschappelijk Onderzoek","keywords":"Computer science; Modality (human–computer interaction); Modular design; Architecture; Computer vision; Artificial intelligence; Sensor fusion; Modalities; Videoconferencing; Multimedia","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":[0.000556777,0.0007222621,0.0004165505,0.0005192905,0.0002901899,0.0005163673,0.0009023005,0.0006854999,0.001438182],"category_scores_gemma":[0.001083918,0.0002706605,0.0003291862,0.0002622338,0.0003641645,0.001132062,0.0008029462,0.0003495837,0.0007079165],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.000268415,"about_ca_system_score_gemma":0.0002464459,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.001018445,"about_ca_topic_score_gemma":0.001199385,"domain_scores_codex":[0.9995456,0.0001483219,0.00001480944,0.00009927605,0.0001327907,0.00005924868],"domain_scores_gemma":[0.9995657,0.0001457913,0.0000550447,0.00005811223,0.0001408386,0.00003450981],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"bench_or_experimental","study_design_gemma":"bench_or_experimental","study_design_scores_codex":[0.0007393722,0.000114988,0.002197803,0.0002475148,0.00009155693,0.0005365168,0.0006369296,0.06804703,0.4926773,0.00828451,0.001181632,0.4252448],"study_design_scores_gemma":[0.00008079468,0.0008983156,0.005210747,0.00005011922,0.0001733016,0.001238359,0.0002735992,0.6200631,0.3557454,0.005694507,0.01044754,0.0001242726],"study_design_candidate":"bench_or_experimental","study_design_consensus":"bench_or_experimental","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.07836221,0.0002374872,0.9186158,0.00005696882,0.00001745611,0.00003206985,0.00002353404,0.00106378,0.001590678],"genre_scores_gemma":[0.7117698,0.0002385896,0.2845116,0.00006502125,0.00004244326,0.00006897063,0.00008203019,0.00006815638,0.003153441],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.001438182,"threshold_uncertainty_score":0.004811168,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.01013697115240418,"score_gpt":0.2277007491503946,"score_spread":0.2175637779979905,"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."}}