CNN Object Detection Models 2

Question 1

What is the primary role of the Region Proposal Network (RPN) in Faster R-CNN?

Accepted Answer

Propose candidate object bounding boxes

Answer

The RPN slides over the feature map and proposes candidate bounding boxes (anchors) that may contain objects.

Question 2

Which loss function does YOLO use for bounding box coordinate regression?

Accepted Answer

Mean squared error (MSE)

Answer

Original YOLO uses MSE for coordinate regression, though later versions switched to IoU-based losses.

Question 3

In SSD (Single Shot MultiBox Detector), what is the purpose of using multiple feature map scales?

Accepted Answer

To detect objects of different sizes

Answer

SSD predicts detections from feature maps at multiple resolutions, allowing it to handle both small and large objects.

Question 4

What does 'anchor-free' mean in the context of object detectors like FCOS?

Accepted Answer

Bounding boxes are predicted without predefined anchor boxes

Answer

Anchor-free detectors like FCOS predict object locations directly from feature map points without relying on predefined anchor boxes.

Question 5

Which metric is typically used to evaluate object detection performance?

Accepted Answer

Mean Average Precision (mAP)

Answer

mAP averages the Average Precision across all object classes and IoU thresholds, making it the standard detection metric.

CNN - Convolutional Neural Networks Practice Test