Help Required for Custom Real Time Object Detection Through Jetson Nano and Pi v2 Camera on my own Dataset

I want to perform real time object detection on my own dataset using Nano and Raspberry Pi v2 camera sensor. I want urgent help as I have a pending project to complete.

Hi farjadhaider3253,

Please refer to:

Thankyou @kayccc , I have seen the link which you have mentioned but that has very exhaustive information. i am actually beginner in this field and secondly I have very less time to complete my FYP that is to perform real time custom object detection on my own dataset. So kindly help me in this regard by telling me the easiest path which I can follow.

@farjadhaider3253, this pages from the tutorial above will get you up & running quickly on the inferencing side:

To train your own detection model on custom dataset, I have been working on new part of the tutorial, it is found in the dev branch of jetson-inference:

I am working on documenting the data collection aspect next. If you already have collected your images, I suggest using a tool like this to label them. It saves the dataset labels in Pascal VOC format, which the PyTorch code above supports reading.

1 Like

Ok @dusty_nv I will definitely check out the tutorial which you have mentioned and will tell you if found some issue. Thanks a lot.