HiwApp: Preliminary Results for an Android Application Culinary Tool Identifying Meat Cuts Using Machine Learning

Abstract

Technological advancements are reshaping the culinary landscape, aiming to improve cooking processes and dining experiences. The integration of machine learning into culinary tools, including meat cut identification, is gaining traction globally. In the Philippines, where culinary heritage is rich and diverse, technology increasingly finds its place in kitchens. Despite this, identifying meat cuts remains challenging, prompting innovative technological solutions. Bohol, a province in the Philippines, reflects this trend, with a growing interest in modern culinary tools among local chefs and home cooks. In this context, the development of HiwApp, an Android application utilizing machine learning for meat cut identification, represents a notable advancement. HiwApp employs supervised machine learning, specifically the k-Nearest Neighbors (k-NN) algorithm, to identify meat cuts that cater to professional chefs and home cooks. This paper introduces HiwApp's development process, detailing its methodology, which includes synthetic data augmentation, algorithm implementation, and user interface design. Preliminary results indicate HiwApp's satisfactory performance, achieving an 84.55% accuracy rate. Future efforts aim to address limitations and enhance HiwApp's meat cut recognition capabilities, improving user culinary experiences. Additionally, recommendations for future development include predicting dishes based on identified cuts, estimating market income, and integrating features for recipe suggestions and freshness prediction to broaden HiwApp's practical applications.