AI-Driven Implicit Consumer Emotion Measurement in Food Sensory Evaluation: Research Progress and Future Directions

Abstract

Traditional food sensory evaluation relies heavily on explicit judgments such as descriptive analysis, check-all-that-apply questions, and hedonic ratings. These methods remain indispensable, yet they are vulnerable to memory limits, social desirability, fatigue, contextual bias, and the difficulty of verbalizing weak or rapidly changing affective states. This review examines how artificial intelligence is accelerating implicit consumer emotion measurement in food sensory evaluation. We synthesize recent progress in facial expression analysis, eye tracking, and electroencephalography, with attention to the affective signals each modality can capture, the signal-processing and machine-learning methods that make those signals usable, and the role of multimodal fusion in more robust preference prediction. The review argues that implicit measurement should be treated as a complementary layer rather than a substitute for explicit sensory science. Its future value depends on calibrated multimodal protocols, transparent models, ecological test environments, and strict ethical governance of biometric data.