Notes

Introduction to Image processing

2 min read

Image processing is the field of where we do operations on an image and get an image

Operations to do

  1. Image Denoising
  2. Image Enhancement - either in Time domain or frequency domain
  3. Edge detection
  4. Image Segmentation
  5. OCR, Optical character recognition
  6. face and smile detection
  7. Biometrics
  8. Object Recognition

Computer vision challenges:

  1. In projection we use depth
  2. View point can change information
  3. Illumination can change information
  4. Occlusion
  5. Scale
  6. deformation
  7. background clutter
  8. Intra class variation

Important notes

Origin of an image starts from the top left,

NOTE: Cross correlation vs convolution is on page 153 Sec 3.4 Gonzalez digital image processing book

Some filters

1 & 1 & 1 \\ 1 & 1 & 1 \\ 1 & 1 & 1 \end{bmatrix}$$ Changing the numbers symmetrically will give a weighted average Weighted average is better because it always yields numbers in the power of two which could be done easily with just `sll` and `slr` instructions Also weighted averages help in being focused on the centered pixels or gives them the higher contribution, which is kind of gives better or more information about what is the current pixel $$weighted \space blurring \space filter (smoothing)= \frac{1}{16}\begin{bmatrix} 1 & 2 & 1 \\ 2 & 4 & 2 \\ 1 & 2 & 1 \end{bmatrix}$$

Graph View