How to look at bbox predictions after training?

[alareza619]

Hi, I finetuned a pretrained Faster RCNN model. I used the instance segmentation Mask RCNN pytorch tutorial as a guide.

I finished training and can't figure out how to look at bbox predictions.

For segmentation prediction, the guide used the following to display the mask Image.fromarray(prediction[0]['masks'][0, 0].mul(255).byte().cpu().numpy())

I tried Image.fromarray(prediction[0]['boxes'].mul(255).byte().cpu().numpy()) But this doesn't work.

link to tutorial I followed : https://colab.research.google.com/github/pytorch/vision/blob/temp-tutorial/tutorials/torchvision_finetuning_instance_segmentation.ipynb#scrollTo=5v5S3bm07SO1

I used the following to train on a custom dataset.

• def get_faster_rcnn_model(num_classes):
• model = torchvision.models.detection.fasterrcnn_resnet50_fpn(pretrained=True)
• num_classes = 2 # 1 class (person) + background
• in_features = model.roi_heads.box_predictor.cls_score.in_features
• model.roi_heads.box_predictor = FastRCNNPredictor(in_features, num_classes)
• return model

[fmassa]

Hi,

I hope you have found how to visualize the boxes. The prediction[0]['boxes'] corresponds to the [x0, y0, x1, y1] coordinates of the pixels in the image, so you can use matplotlib.pyplot.Rectangle to plot it