1. we have an existing program that runs on Linux; it captures videos and then saves images from the video and uploads to a cloud server. during very busy days, when we have more people walking into the stores through the front doors it is missing some images when it uploads the program uses a camera to record videos mounted in front of the front doors meaning the video frames are freezing and so if people are walking in, it misses those while it is frozen the program is written in C, C++ and Python and uses openCV for image recognition. It runs on a Linux Ubuntu OS with 1 or more camera mounted in front of the front door.
2. we have a few stores with more than one camera and i need a change in the program so i know which camera reported that image; right now all images come into one folder and then uploads Bottom line - we are only able to count 90% of all people coming in and not a 100% - we want to be at 100% counting for all "faces" only coming in - maybe clear faces or not so clear but definitely not counting "non-face" images.
- Must have experiences with OpenCV.
- Must have experiences with Nvidia cuDNN Library
- Must have experiences with Cuda Toolkit
- Must have experiences with C++/Bash Shell/Python
- Must have IP cameras.
47 pekerja bebas membida secara purata $2470 untuk pekerjaan ini
Hi, have experience in face recognition and M.L using python and opencv. We can look into your system and can work to improvise it. Lets discuss over the chat to discuss more.
Hi there, I have checked the details I have great experience with C++ Programming, Face Recognition, Machine Learning, OpenCV, Python. Please start the chat so we can discuss this job more in detail. Thanks