HALCON's Deep-Learning-Based Object Detection 3: Evaluate the Trained Model

In the third part of this tutorial series on HALCON Object Detection, we will evaluate the deep-learning-based object detection model we trained in the previous video. First, we will have a look at how you can influence the inference results by filtering them according to their overlaps and assigned confidence values. Then, we will inspect some pie charts, visualizing the precision and recall of the model on a given dataset. Additionally, a confusion matrix helps us analyze the classes in detail. Lastly, we show how to visualize images with false positives and false negatives. 0:45 – Visualize some results 1:22 – Adapt the inference output with parameters 2:35 – Evaluate the pie charts for precision and recall 3:45 – Analyze the confusion matrix 4:22 – Visualize false positives and false negatives In this video, HALCON 19.11 is used. www.mvtec.com www.halcon.com