Individual calibration improves facial emotion recognition tasks
Toward individual-level calibration in affect recognition with perceptual adjustment queries
Machine Learning
Summary
People don't all perceive facial emotions with the same sensitivity, which makes standard tests assume everyone finds these tasks equally hard, even when they don't. The authors tested a method that measures how small a change in facial emotion each person can notice, using easy questions. They then use this measure to adjust task difficulty so it's fair for each participant. This approach made task difficulty feel more equal for individuals, and it helped make people respond faster and more consistently.
What this means in practice
- •For human-computer interaction designers: Improve emotion recognition assessments by customizing difficulty to each user's perceptual sensitivity, leading to more accurate and fair results.
- •For psychological testing services: Offer perceptually calibrated tests that reduce variability and response times in emotional recognition tasks across clients.$Commercial implications: Enables development of individualized diagnostic tools for emotion perception-based assessments, improving service precision and scalability.
Authors
Xuanzhou Chen, Sankaraleengam Alagapan, Ashwin Pananjady
Abstract
Behavioral tasks measuring facial affect perception assume that identical stimuli impose equivalent perceptual difficulty across participants. However, this assumption is systematically violated by individual differences in perceptual sensitivity. Using an affective perception task as our testbed, we propose a framework to normalize for perceptual difficulty that directly estimates each participant's Just Noticeable Difference (JND) along the facial affect spectrum via cognitively lightweight perceptual adjustment queries (PAQs). We use these PAQ-inferred JNDs to re-express stimulus distances, constructing difficulty-equated tasks in perceptual space. We validate the framework in a Two-Alternative Forced-Choice (2AFC) task using two complementary behavioral measures: binary metacognitive difficulty judgments and response time variance decomposition. We find that PAQ calibration significantly equalizes perceived task difficulty at an individual level when compared to both the non-calibrated baseline and population-level Weibull calibration, while also reducing mean response time and between-subject variance in response time. These results establish PAQ as a principled and practical instrument for individualized perceptual calibration in facial affect recognition.