A Deep Learning-Based Rotten Food Recognition App for Older Adults: Development and Usability Study

Area of Science:

• Gerontology
• Computer Science
• Food Science

Background:

• Cognitive decline in older adults increases the risk of food poisoning from consuming rotten produce.
• Existing tools for detecting rotten food lack specific applications for older adults.
• A need exists for user-friendly technology to support older adults in identifying unsafe food items.

Purpose of the Study:

• To develop a smartphone application for older adults to identify rotten fruits via image classification.
• To assess the usability and user perceptions of the developed food freshness detection app among older adults.

Main Methods:

• A smartphone app was created utilizing residual deep networks for fruit freshness analysis.
• Healthy older adults (aged 65+) participated in usability testing and provided feedback through surveys and interviews.
• App performance was evaluated using after-scenario questionnaires and qualitative interview data.

Main Results:

• Participants found the app simple, easy to use, and visually appealing for determining fruit freshness.
• Older adults were satisfied with the app's efficiency in identifying fresh fruit but hesitant towards a paid version.
• Qualitative feedback indicated no difficulties in capturing fruit images with the app.

Conclusions:

• The developed app demonstrates potential for effectively assisting older adults in identifying rotten food.
• Future research should expand the app's capability to detect the freshness of a wider variety of food items.