Project

AI Diet & Exercise Recommendation System (BSCS Capstone)

Independent (NUCES-FAST)

Full-stack fitness platform delivering personalized diet and exercise regimens powered by Mask R-CNN muscle assessment and Apriori recommendation algorithms.

LiveIndependent (NUCES-FAST)1 Aug 2020 — 30 Jul 2021 · 11 mo

PythonPyTorchMask R-CNNApriori AlgorithmNATS StreamingDockerKubernetesReact NativeNode.jsExpress.jsMongoDBNginx

Engineered as the Senior Final Year Capstone Project at NUCES-FAST, this platform automates personalized diet and workout regimens by first conducting an intelligent routine and visual body assessment. A computer vision pipeline segments and analyzes muscle groups using Mask R-CNN, while an Apriori association rule-mining engine personalizes macro-nutritional diets and exercise progressions based on individual fitness baselines.

Architected on an event-driven microservices infrastructure with NATS Streaming, Docker, and Kubernetes, the system decouples Express.js client-facing APIs (User, Exercise, Nutrition, and Tracking services) from asynchronous Python AI inference workers. A cross-platform React Native mobile client provides interactive macro tracking and routine guidance.

The core computer vision muscle assessment subsystem was subsequently presented at the 4th IEEE International Conference on Computing & Information Sciences (ICCIS 2021) and published in IEEE Xplore.

Problem

Generic fitness apps provide static, one-size-fits-all workout and diet plans that fail to adapt to a user's actual muscular condition, physical baseline, or dietary habits, while personal trainers remain expensive.

Approach

Built an event-driven microservices architecture using NATS Streaming, Docker, and Kubernetes. Deployed Python services leveraging Mask R-CNN for visual muscle assessment and Apriori rule mining for diet scheduling, integrated with a React Native mobile application and Express.js backends.

Results

Awarded Grade A+ for the Senior Capstone Project. The underlying visual muscle assessment engine was peer-reviewed, presented at IEEE ICCIS 2021, and published in IEEE Xplore.

Diet and Exercise Recommendation System Architecture Diagram and Mobile App — Diet & Exercise Recommendation System Architecture & Mobile AppOfficial BSCS Final Year Capstone overview: React Native mobile interface, NATS event-driven microservices topology, and tech stack (NUCES-FAST).

Computer vision muscle assessment service published in IEEE Xplore
Presented at the 4th IEEE International Conference on Computing & Information Sciences (ICCIS 2021)
IEEE Xplore Document: 9676387

Published IEEE Research Paper

Connected work

Skills

Publications

A Convolutional Neural Network-based Framework for the Assessment of Human Muscles

Education

BS in Computer Science — NUCES-FAST

Loading content…

The core computer vision muscle assessment subsystem was subsequently presented at the 4th IEEE International Conference on Computing & Information Sciences (ICCIS 2021) and published in IEEE Xplore.