{"id":"W4415649506","doi":"10.1111/1556-4029.70203","title":"Learning to localize image forgery using boundary‐preserving mask R‐ <scp>CNN</scp>","year":2025,"lang":"en","type":"article","venue":"Journal of Forensic Sciences","topic":"Digital Media Forensic Detection","field":"Computer Science","cited_by":0,"is_retracted":false,"has_abstract":true,"ca_institutions":"Kellogg's (Canada)","funders":"","keywords":"Preprocessor; Benchmark (surveying); Compositing; Segmentation; Domain (mathematical analysis); Image (mathematics); Digital forensics; Channel (broadcasting)","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":["scholarly_communication"],"consensus_categories":[],"category_scores_codex":[0.002093044,0.0002005637,0.0003673197,0.001002979,0.0005115033,0.001388616,0.001541097,0.0000670655,0.000004377454],"category_scores_gemma":[0.003183996,0.0001642425,0.0001953416,0.002746979,0.0004848026,0.003093521,0.0005803301,0.0003485009,0.00001833498],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0001712247,"about_ca_system_score_gemma":0.0006616834,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00003082489,"about_ca_topic_score_gemma":0.00001562888,"domain_scores_codex":[0.9971886,0.00009622496,0.0006714294,0.0003884734,0.001076211,0.000579097],"domain_scores_gemma":[0.9978731,0.000640302,0.0004841726,0.0003027875,0.0004605876,0.0002390714],"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":[0.00003966767,0.0002057985,0.01132878,0.0001721083,0.000212623,0.0004856314,0.00816641,0.03539361,0.04621291,0.01567104,0.04488413,0.8372273],"study_design_scores_gemma":[0.002089883,0.004239548,0.009773623,0.003380604,0.0001644931,0.0021667,0.01265931,0.4506903,0.2207419,0.1871447,0.1061375,0.0008114261],"study_design_candidate":"design_other","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.6301383,0.0002562656,0.3537863,0.0008684254,0.004219137,0.000117904,4.81463e-7,0.00006528259,0.01054787],"genre_scores_gemma":[0.813565,0.00001406836,0.1844937,0.0007580377,0.0002954419,0.00000231477,1.627568e-7,0.00001172439,0.0008595761],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.8364159,"threshold_uncertainty_score":0.999648,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.01779093602414762,"score_gpt":0.2804418975105068,"score_spread":0.2626509614863592,"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."}}