Skip to content
Running object detection on a webcam feed using TensorRT on NVIDIA GPUs in Python.
Python Dockerfile Shell
Branch: master
Clone or download
Latest commit feb3632 Jun 19, 2019
Type Name Latest commit message Commit time
Failed to load latest commit information.
SSD_Model Adding code Jun 19, 2019
dockerfiles Adding code Jun 19, 2019 Update Jun 19, 2019 Adding code Jun 19, 2019

Object Detection TensorRT Example:

This python application takes frames from a live video stream and perform object detection on GPUs. We use a pre-trained Single Shot Detection (SSD) model with Inception V2, apply TensorRT’s optimizations, generate a runtime for our GPU, and then perform inference on the video feed to get labels and bounding boxes. The application then annotates the original frames with these bounding boxes and class labels. The resulting video feed has bounding box predictions from our object detection network overlaid on it. The same approach can be extended to other tasks such as classification and segmentation.

A detailed explanation of this code sample can be found in the How to Apply Deep Learning for Common Applications webinar and also as a blog on the NVIDIA Medium Page.

1. Setup the environment


2. Run inference on the webcam

python SSD_Model/ 

Note: By changing the argument to the "p" flag, you can change which precision the model will run in. The options are FP32 (-p 32), FP16 (-p 16), and INT8 (-p 8)

Note: When running in INT8 precision, an extra step will be performed for calibration. But just like building an engine, it will only be performed the first time you run the model.

You can’t perform that action at this time.