Performance measures for object detection
5 important questions on Performance measures for object detection
Give two reasons why object classification is more complex than image classification.
2. Label has to be matched on its presence in the image.
What is the use of an overlapping threshold in evaluating the performance of an object detection?
How is the algorithm called to compute the overlapping threshold between two bounding boxes? How to compute it?
- Higher grades + faster learning
- Never study anything twice
- 100% sure, 100% understanding
What is the algorithm called to filter out duplicated bounding boxes (that identify the same object) in object detection? How does it work?
1. First it sorts the list of bounding boxes on their respective confidence score.
2. Iteratively remove bounding boxes that overlap with those of higher confidence score.
Name three performance measures for object detection.
2. False positive per window: Less false positives require less computation power.
3. False positive per image: Same as 2.
The question on the page originate from the summary of the following study material:
- A unique study and practice tool
- Never study anything twice again
- Get the grades you hope for
- 100% sure, 100% understanding