Evaluation

These exercises will introduce you to different metrics used for evaluating the performance of a classifer. Additionally, you will also implement a face classifier using HOG features.

Learning goals

After this week's exercises, you should be able to:

  • Deliberate appropiate evaluation metrics for classifiers.
  • Reason about and construct the confusion matrix for classification evaluation.
  • Inspect the metrics and reflect on their significance related to the charateristics of the dataset.
  • Implement and experiment with HoG features for classification.
  • Reason about using HoG features as classifier.
  • Reason about HoG features and sliding window for classification of images.