Automated Action Verification for Schizophrenia Cognitive Remediation Using Vision-Language Models
A research paper on arXiv (2607.22721) proposes an automated framework using Vision-Language Models to verify actions in cognitive remediation tasks for schizophrenia patients. The system uses a camera-monitored tabletop with miniature scenes (roads, roundabout, park, toy vehicles). Patients receive audio instructions for goal-oriented spatial actions with a toy vehicle. The approach aims to reduce subjectivity and improve scalability of manual clinician observation.
Key facts
- arXiv paper ID: 2607.22721
- Proposes automated action verification for cognitive remediation
- Uses Vision-Language Models
- Targets schizophrenia rehabilitation
- Tabletop environment with miniature scenes: roads, roundabout, park, toy vehicles
- Patients manipulate toy vehicle based on audio instructions
- Aims to reduce subjectivity of manual observation
- Published on arXiv.org
Entities
Institutions
- arXiv