2020/06/10 by Ali-Akbar Samadani, Samadani, Ali, Rob Gorbet +4 · 1 citation
Computer Science · Psychology · #Emotion and Mood Recognition #FOS: Computer and information sciences #Gaze Tracking and Assistive Technology #Graphics (cs.GR) #Hand Gesture Recognition Systems #Human Pose and Action Recognition #Human-Computer Interaction (cs.HC) #Machine Learning (cs.LG)
paper · pdf · doi:10.48550/arxiv.2006.06071
openalex publication_date 2020/06/10 · openalex created_date 2022/07/26 · openalex updated_date 2026/07/28
Body movements are an important communication medium through which affective\nstates can be discerned. Movements that convey affect can also give machines\nlife-like attributes and help to create a more engaging human-machine\ninteraction. This paper presents an approach for automatic affective movement\ngeneration that makes use of two movement abstractions: 1) Laban movement\nanalysis (LMA), and 2) hidden Markov modeling. The LMA provides a systematic\ntool for an abstract representation of the kinematic and expressive\ncharacteristics of movements. Given a desired motion path on which a target\nemotion is to be overlaid, the proposed approach searches a labeled dataset in\nthe LMA Effort and Shape space for similar movements to the desired motion path\nthat convey the target emotion. An HMM abstraction of the identified movements\nis obtained and used with the desired motion path to generate a novel movement\nthat is a modulated version of the desired motion path that conveys the target\nemotion. The extent of modulation can be varied, trading-off between kinematic\nand affective constraints in the generated movement. The proposed approach is\ntested using a full-body movement dataset. The efficacy of the proposed\napproach in generating movements with recognizable target emotions is assessed\nusing a validated automatic recognition model and a user study. The target\nemotions were correctly recognized from the generated movements at a rate of\n72% using the recognition model. Furthermore, participants in the user study\nwere able to correctly perceive the target emotions from a sample of generated\nmovements, although some cases of confusion were also observed.\n