Core Concepts
What type of Ai task is computer vision mainly concerned with?
Visual perception and interpretation
Which algorithm detects edges in images?
Canny edge detector
What type of learning is commonly used in computer vision?
Supervised learning
What vision task is used in autonomous driving for lane detection?
Image segmentation
What does CNN stand for?
Convolutional Neural Network
What is the difference between image processing and computer vision?
Image processing enhances images, while computer vision understands and interprets them
Which neural network layer is most important for feature extraction?
Convolutional layer
What is overfitting in vision models?
When a model performs well on training data but poorly on new dataManual annotation is time-consuming and requires expertise
Why is real-time processing critical in self-driving cars?
Delays can cause unsafe decisions
What is transfer learning in computer vision?
Using a pre-trained model for a new task
Why is computer vision considered a perception problem?
Because it converts raw visual data into meaningful understanding
What is the purpose of pooling layers in CNNs?
To reduce spatial dimensions and computation
Why is labeled data expensive in computer vision?
Manual annotation is time-consuming and requires expertise
What is a major challenge in medical image analysis?
High accuracy requirement and risk of errors
What is the difference between object detection and segmentation?
Detection finds objects; segmentation defines exact object boundaries
What does “semantic understanding” mean in computer vision?
Assigning meaning and labels to visual content
Name one object detection algorithm.
YOLO / Faster R-CNN / SSD
What is data augmentation?
Artificially increasing data using transformations like rotation or flipping
Why does facial recognition raise ethical concerns?
Privacy invasion and bias
Why are CNNs translation invariant?
Because shared weights detect features anywhere in the image
Why is computer vision difficult compared to human vision?
Visual scenes are complex, ambiguous, and vary in lighting, angle, and scale
Why are Vision Transformers challenging compared to CNNs?
They require large datasets and high computational power
What problem occurs when training data lacks diversity?
Bias and poor generalization
Why is computer vision unreliable in low-light environments?
Poor image quality and loss of features
What is the biggest limitation of current computer vision systems?
Lack of true human-level understanding and reasoning