Facial expression analysis with AFFDEX and FACET: A validation study
[FULL TEXT READ at the publisher. Peer-reviewed, Psychonomic Society] Stockli, S., Schulte-Mecklenbeck, M., Borer, S., & Samson, A. C. (2018). 'Facial expression analysis with AFFDEX and FACET: A validation study.' Behavior Research Methods, 50(4), 1446-1460. doi:10.3758/s13428-017-0996-1. THE SINGLE BEST SET OF NUMBERS IN THE WHOLE FILE AND THE ARTICLE SHOULD BUILD A PARAGRAPH ON IT. STUDY 1, standardised posed photographs from the Amsterdam Dynamic Facial Expression Set, the Warsaw Set of Emotional Facial Expression Pictures and the Radboud Faces Database, matching score, non-baseline-corrected: FACET 97 PER CENT OVERALL, being 99 per cent on Amsterdam, 92 per cent on Warsaw and 99 per cent on Radboud; AFFDEX 73 PER CENT OVERALL, being 73, 66 and 77 per cent. Baseline-corrected these become 95 and 72 per cent. AFFDEX failed to detect a face at all in 1 per cent of pictures; FACET had no detection failures. AFFDEX's weakest categories were anger and fear, FACET's sadness and fear; AFFDEX 'usually confused fear with surprise' and 'often confused anger with sadness'. STUDY 2, real participants: 119 students at a Swiss university recruited, FINAL SAMPLE 110 (63 female, mean age 21.20), nine excluded for detection failure, and A METHODOLOGICALLY CRITICAL SENTENCE, VERBATIM: 'Only Caucasian participants without facial artifacts (e.g., disruptive glasses, beards, or scarves) were included.' Stimuli were emotionally evocative pictures from the International Affective Picture System and the Geneva Affective Picture Database for a valence task, plus imitation of Radboud portraits. VALENCE CLASSIFICATION, baseline-corrected: AFFDEX 57 PER CENT OVERALL, being 17 PER CENT correct on responses to positive pictures and 97 per cent on negative; FACET 67 PER CENT OVERALL, 63 per cent positive and 71 per cent negative. Picture-wise variation was large: for one Geneva picture, AFFDEX
[FULL TEXT READ at the publisher. Peer-reviewed, Psychonomic Society] Stockli, S., Schulte-Mecklenbeck, M., Borer, S., & Samson, A. C. (2018). 'Facial expression analysis with AFFDEX and FACET: A validation study.' Behavior Research Methods, 50(4), 1446-1460. doi:10.3758/s13428-017-0996-1. THE SINGLE BEST SET OF NUMBERS IN THE WHOLE FILE AND THE ARTICLE SHOULD BUILD A PARAGRAPH ON IT. STUDY 1, standardised posed photographs from the Amsterdam Dynamic Facial Expression Set, the Warsaw Set of Emotional Facial Expression Pictures and the Radboud Faces Database, matching score, non-baseline-corrected: FACET 97 PER CENT OVERALL, being 99 per cent on Amsterdam, 92 per cent on Warsaw and 99 per cent on Radboud; AFFDEX 73 PER CENT OVERALL, being 73, 66 and 77 per cent. Baseline-corrected these become 95 and 72 per cent. AFFDEX failed to detect a face at all in 1 per cent of pictures; FACET had no detection failures. AFFDEX's weakest categories were anger and fear, FACET's sadness and fear; AFFDEX 'usually confused fear with surprise' and 'often confused anger with sadness'. STUDY 2, real participants: 119 students at a Swiss university recruited, FINAL SAMPLE 110 (63 female, mean age 21.20), nine excluded for detection failure, and A METHODOLOGICALLY CRITICAL SENTENCE, VERBATIM: 'Only Caucasian participants without facial artifacts (e.g., disruptive glasses, beards, or scarves) were included.' Stimuli were emotionally evocative pictures from the International Affective Picture System and the Geneva Affective Picture Database for a valence task, plus imitation of Radboud portraits. VALENCE CLASSIFICATION, baseline-corrected: AFFDEX 57 PER CENT OVERALL, being 17 PER CENT correct on responses to positive pictures and 97 per cent on negative; FACET 67 PER CENT OVERALL, 63 per cent positive and 71 per cent negative. Picture-wise variation was large: for one Geneva picture, AFFDEX
Written from the fULL TEXT READ at the publisher. Peer-reviewed, Psychonomic Society.
What our sources record
Recorded when this source was checked, figures included, with page references and quotation marks removed.
Stockli, S., Schulte-Mecklenbeck, M., Borer, S., & Samson, A. C. (2018). Facial expression analysis with AFFDEX and FACET: A validation study. Behavior Research Methods, 50(4), 1446-1460. doi:10.3758/s13428-017-0996-1