Strength Training: A Fitness Application for Indoor Based Exercise Recognition and Comfort Analysis

Mobile and Wearable health applications are playing a key role in monitoring user fitness data. Current Wearable devices like Fitbit, Jawbone, and Samsung Gear provide solutions for outdoor fitness activities. However a lot of people prefer to do indoor exercises but there are limited auto activity tracking solutions. Existing solutions are not concentrated on the ease at which a user is able to perform an exercise. This work provides automatic indoor exercise recognition for both in gym and home usage scenarios. Activities under consideration are Biceps curl, Chest fly, Row, Push up, Sit up, Squat and Triceps curl. Accuracy of 95.3% and 99.4% is achieved for activity recognition and repetition count respectively. Along with activity recognition this work targets at analyzing comfort and measuring calorie burnt during the exercise. We introduce Comfort factor, which is the state of physical ease during workout. Comfort factor is predicted using the relation devised between Resting Heart Rate (RHR), Maximum Heart Rate (MHR), Real-time heart rate value and hand tremor at exercise exhaustion limit of the user. With this comfort factor, user can decide whether to go for more weights or reduce weights in weight training activities.