Name: Catherine Gai

SID: 3034712396

Email: catherine_gai@berkeley.edu

Link to project report website: file:///Users/catherine_gai/Desktop/Cal/2021%20Fall/CS%20194-26/proj1/deliverables/%20Project%2001.html

This folder contains four functional python files: "single_scale.py", "multi_scale.py", "utils.py" and "main.py".

**single_scale.py: **

This python file contains all functions required to run the naive exhaustive approach for aligning channels, as well as automatic contrast and automatic cropping after alignment.

• single_scale_multi(params): exhaustive search with normalized cross correlation metric, used during single iteration in the image pyramid search.

• single_scale(params): exhaustive search with l2-norm metric. Used during naive exhaustive search.

• l2_norm(params): calculates the l2-norm between two images.

• ncc(params): calculates the normalized cross correlation between two images.

• generate(params): generates synthesized images for all input JPEG low-resolution files.

• main(): set up directory, collects images, and calls generate. Also called by other programs to run "single_scale.py".

**multi_scale.py: **

This python file contains all functions required to run the image pyramid approach for aligning channels, as well as automatic contrast and automatic cropping after alignment.

• multi_scale(params): executes image pyramid search on given R, G and B channels. Need to specify the number of layers of the pyramid.
• generate(params): generates synthesized images for all input TIFF low-resolution files.
• main(): set up directory, collects images, and calls generate. Also called by other programs to run pyramid search on required 11 images.
• extra(): set up directory, collects images, and calls generate. Also called by other programs to run pyramid search on additional 2 images.

utils.py:

This python file contains all functions for Bells and Whistles part. Two main portions are included here: automatic contrast and automatic cropping.

• auto_contrast_lab(params): provided an RGB image, first transforms it into LAB format and do histogram equalization on the L (luminance) layer.
• auto_contrast_rgb(params): provided an RGB image, do histogram equalization on all three layers respectively.
• auto_cropping(params): provided an RGB image, crop out the boundaries that has only one or two color channels.

main.py:

Run this python file to execute naive exhaustive search or image pyramid method. Three arguments can be specified to control execution of naive exhaustive search (on 3 required JPEG images), image pyramid (on 11 required TIFF images) and extra (2 additional TIFF images). Arguments are specified below:

• --single_scale (or -s):
  • Takes on two values "True" and "False" (note that both are string format).
  • Default value: "True".
  • If "True", runs the naive exhaustive search on 3 required JPEG images.
• --multi_scale (or -m):
  • Takes on two values "True" and "False" (note that both are string format).
  • Default value: "True"
  • If "True", runs the image pyramid method on 11 required TIFF images.
• --extra (or -e):
  • Takes on two values "True" and "False" (not that both are string format).
  • Default value: "True"
  • If "True", runs the image pyramid method on 2 additional TIFF images.

Sample execution commands:

• python main.py -s="True" -m="False" -e="False": only performs exhaustive search on 3 required JPEG images.
• python main.py -m="False": performs exhaustive search on 3 required JPEG images and image pyramid method on 2 additional TIFF images.
• python main.py: performs all executions.