| 英文摘要 |
This study aims to explore the application of artificial intelligence (AI) technology in emotion recognition within portraiture and analyze its connection to formalism theory. Portraits have long served not only to depict physical likeness but also to express emotions and convey sentiments. Emotion recognition, a key AI technology, can reveal the hidden emotional layers in portraits when applied to visual art. Using twelve portraits by Taiwanese artist Chiu Ya-Tsai as examples, this research delves into the diversity of emotional expression through AI analysis, focusing on eight primary emotional categories: anger, contempt, disgust, fear, happiness, neutral, sadness, and surprise. The study examines four key elements: facial expressions, body language, attire, and contextual atmosphere. Two AI emotion recognition techniques are compared: Technique A, which focuses on ''facial recognition'' and evaluates emotions solely through micro-expressions, yielding less significant results and exhibiting limitations in its recognition capacity; and Technique B, which adopts a more comprehensive ''holistic recognition'' approach, analyzing not only facial expressions but also body language, attire, and contextual factors. The latter shows higher precision in capturing richer emotional layers beyond facial expressions. Emotional expression in portraits cannot rely solely on facial cues but requires consideration of various visual elements for more accurate emotion interpretation. From the perspective of analyzing emotions through visual features, AI technology and formalism share similarities, as both recognize emotional expression or resonance in art through formal elements. Therefore, the results of emotion recognition in portraiture not only broaden the aesthetic perspective in art but also reinforce the practical application of formalism theory in contemporary art. |