Advanced Deep Learning with Python

Implementing Faster R-CNN with PyTorch

In this section, we'll use a pretrained PyTorch Faster R-CNN with a ResNet50 backbone for object detection. This example requires PyTorch 1.3.1, torchvision 0.4.2, and python-opencv 4.1.1:

1. We'll start with the imports:

import os.pathimport cv2import numpy as npimport requestsimport torchvisionimport torchvision.transforms as transforms

2. Next, we'll continue with downloading the input image, and we'll define the class names in the COCO dataset. This step is the same as the one we implemented in the A code example of YOLOv3 with OpenCV section. The path to the download image is stored in the image_file = 'source_2.png' variable, and the class names are stored in the classes list. This implementation ...