Identification of Cooking ADL Actions through Analysis of Thermal Camera Video

Kitchen use and nutrition are fundamental to well-being and independence for aging adults. Good nutrition enables healthy aging and is necessary for good health, while poor nutrition can lead to health decline. These declines can themselves lead to a reduced meal preparation ability or complexity resulting in continued poor nutrition. Meal preparation is an important Activity of Daily Living (ADL), and this work proposes a method to analyze thermal video for stove top cooking with a method to identify flip events within meal preparation. Flip events are unique to the cooking of some food types and identification allows this cooking behaviour to be assessed. The method applies image processing techniques to the thermal video to identify the food items within the pan and proposes a method to identify flips. A balanced set of cooking recordings that include flip and non-flip events were used and the method provides a sensitivity of 100% and a positive predictive value of 74%. The method can augment kitchen activity of daily living monitoring of cooking habits of aging adults and can assist care givers and clinicians in assessing the changes in the older adult.