Machine Learning Computer Vision 4

Question 1

Which network architecture is specifically designed for real-time object detection with a single forward pass?

Accepted Answer

YOLO (You Only Look Once)

Answer

YOLO performs detection in a single network pass by dividing the image into a grid and predicting boxes and classes simultaneously, enabling real-time performance.

Question 2

What is the function of the encoder in an encoder-decoder architecture like U-Net?

Accepted Answer

Progressively downsampling input to capture high-level semantic features

Answer

The encoder uses convolutional and pooling layers to reduce spatial dimensions while increasing feature depth, capturing abstract semantic representations.

Question 3

What is optical flow used for in computer vision?

Accepted Answer

Estimating the motion of objects between consecutive video frames

Answer

Optical flow estimates apparent pixel-level motion between consecutive frames, capturing the velocity field of moving objects in video.

Question 4

Which technique allows a CNN trained on ImageNet to be adapted for a smaller domain-specific dataset with limited data?

Accepted Answer

Transfer learning

Answer

Transfer learning reuses weights from a model pre-trained on a large dataset, fine-tuning them on the target task to achieve good performance with limited data.

Question 5

In the context of image recognition, what does 'top-5 accuracy' mean?

Accepted Answer

The fraction of samples where the true label appears in the model's five highest-probability predictions

Answer

Top-5 accuracy counts a prediction as correct if the ground-truth class is among the five classes assigned the highest probabilities by the model.

Machine Learning Practice Test

Machine Learning Practice Test

Machine Learning Computer Vision 4