{"id":"W4403096679","doi":"10.1145/3698399","title":"Motion-Aware Self-Supervised RGBT Tracking with Multi-Modality Hierarchical Transformers","year":2024,"lang":"en","type":"article","venue":"ACM Transactions on Multimedia Computing Communications and Applications","topic":"Video Surveillance and Tracking Methods","field":"Computer Science","cited_by":7,"is_retracted":false,"has_abstract":true,"ca_institutions":"University of Ottawa","funders":"National Natural Science Foundation of China","keywords":"Computer science; Modality (human–computer interaction); Artificial intelligence; Computer vision; Transformer; Human–computer interaction","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":[0.0005606248,0.0005641175,0.000467955,0.0004022954,0.0002241048,0.0005953562,0.001013522,0.000459112,0.001574357],"category_scores_gemma":[0.001313794,0.0003250941,0.0005803235,0.0004619052,0.0003844324,0.001369902,0.001194102,0.0006157461,0.0005618666],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0003538658,"about_ca_system_score_gemma":0.0005301241,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.002009656,"about_ca_topic_score_gemma":0.002918533,"domain_scores_codex":[0.9997067,0.00004456252,0.00001402364,0.0001010046,0.00009963865,0.00003400331],"domain_scores_gemma":[0.9996939,0.00006777658,0.00004631227,0.00007333507,0.00009344305,0.00002528861],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"bench_or_experimental","study_design_scores_codex":[0.0003308735,0.0001159222,0.002159567,0.0001775215,0.0001074154,0.000141767,0.0002608317,0.1050718,0.1940531,0.01256021,0.003743566,0.6812774],"study_design_scores_gemma":[0.00001479685,0.00006653853,0.0008989318,0.000009739988,0.00002962241,0.0001304675,0.0000206646,0.9589704,0.03301328,0.004313338,0.002511812,0.00002048299],"study_design_candidate":"bench_or_experimental","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.01064372,0.0001007273,0.9875841,0.00003168602,0.00002092872,0.00002238837,0.00003191511,0.0006775493,0.0008870231],"genre_scores_gemma":[0.5827905,0.0003335967,0.4111451,0.0002054435,0.00005882991,0.0001185166,0.0002876013,0.0002501706,0.004810159],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.002009656,"threshold_uncertainty_score":0.005266786,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.04030909776654987,"score_gpt":0.3265042370937342,"score_spread":0.2861951393271843,"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."}}