1. Q1: A dataset contains 1,000 images of cats and 1,000 images of dogs.
Describe how
you would design a neural network to classify the images, including the architecture,
loss function, and evaluation metrics. Solve this
2. Discuss how CNNs automatically learn features compared to manually designed
features in earlier Computer Vision approaches. Include examples.
3. An image is rotated by 45°. Write the transformation matrix and explain its effect on
the image coordinates.
4. Numerical on Scaling the geometry coordinates.
5. Derive the expression for PCA and Explain the role of eigenvalues and eigenvectors in
Principal Component Analysis (PCA) for dimensionality reduction in image processing.
6. Explain the concepts of Image Formation and why do we always consider RGB only.
7. What is Fisher Linear discriminant Analysis, derive the mathematical expression for
the same
8.
9. Numerical on Confusion Matrix
10. Numerical on Bayes theorem.
11. In medical imaging, you are tasked to detect tumors in MRI scans. What preprocessing
and deep learning methods would you use?
12. What do you mean by Image matting and compositing, Texture analysis and synthesis
13. Design principles of pattern recognition system, Learning and adaptation, Pattern
recognition approaches
14. Pseudo codes of RCNN, CNN, Architecture of CNN, VGG16
15.
16. How do you think DSLR camera differ from Mobile camera.
17. Image Segmentation Using Gaussian Mixture Models (GMMs)
18. Gabor filtered related questions
19. SVM derivation and MATHEMATICAL
20. PDF, CDF, MEAN, EXPECTATION, VARIANCE, COVARIANCE CONCEPTS AND formulas
21. How to calculate w, projection vector