We are trying to make a store visitors counter. Where with machine learning technics visitors are counted. The camera is recording people from the top view, and the objective is to do the processing on the Raspberry Pie and not on a server. We have implemented tiny-Yolo for objected detection, but we have some problems loading the darknet Yolo model. The problem is that the Pi3 B+ has limited resources. Tracking is taking a lot of resources and for improving the speed we are working on making multitracking from all four cores of raspberry pi separately. This seems to work now, but we still have some issues with the pipeline.
We need someone with good machine learning and python skills to help our current team to finish the work they are working on now. So this project is more checking the current code structures and advising the current team with ideas how to make the solution work. So the goal of this project is not to remake our current solution, but to help to get our current code working on the raspberry Pie with opencv, yolo and other machine learning technics without having the processing on the server.
12 freelancers are bidding on average $12/hour for this job
Hi I have been working on google vision kit with rasberry pi. Below is demo [login to view URL] We can have a small discussion, So i can share complete break down for this task.
I have made projects like this and have prior experience in the modules which are to be used in the projects. I will complete this project on time as said. So please provide us with this project.