I've developed a Python script that uses libraries like OpenCV, Ultralytics YOLO, using yolo's pose estimation model to detect the no. of steps taken by a human from a video feed( a device's camera).
Although, It works good on my mac m1 , I'm looking to convert this Python code into an Android app so that I can run it on my smartphone. I'm very new to android development. What approaches/ideas can I follow to develop such kind of an mobile application?
Here are the main challenges I'm encountering: Camera Integration: The Python code uses OpenCV to capture frames from the webcam. I need to replace this with the Android Camera API to capture frames from the device's camera. I'm not sure how to properly integrate this API into the Android app. Object Detection: The Python code utilizes Ultralytics YOLO for object detection. Unfortunately, this library is not directly compatible with Android. I need to find an alternative object detection library that works on Android and integrate it into the app.
Or can I use a python backend API which calls my python script to the frontend android script like Java/kotlin? I'm doubting this because it might introduce latency issues and I really want it in real-time. I'm completely lost on this. Please help.
I've tried using kivy but it does not run for some reason.