Culinary Captures: A First Step Towards Fine-Grained Cooking Activity Summarization from Video to Guide Clinicians’ Intervention Plans

The growing elderly population creates a significant challenge for healthcare systems. Assessing a patient’s ability to perform daily activities such as meal preparation is crucial for tailoring care plans. This paper explores the potential of a video analysis system using computer vision to generate reports on cooking activities. This system extracts information on ingredients, utensils, and actions performed, offering a more objective and comprehensive assessment compared to traditional methods. While the current iteration focuses on core information, the system inherently possesses the potential to analyze the temporal sequence of actions. Future work will refine object detection and action recognition to improve accuracy and unlock the full potential of temporal analysis.