adityashrm21 / image-compression-toolkit-R

Three ways to compress your images!

Image Compression Toolkit - R

Three ways to compress your images!

Contributors

• Aditya Sharma
• Alden Chen
• Sayanti Ghosh

Project Summary

This R package specializes in reducing the size of images. It contains three main functions :seam_carve(), crop(), and compress(). The seam_carve() and crop() functions reduce the size of an image by reducing the height and width of an image to the size specified by the user. The compress() function reduces the size of an image by reducing the number of bits used in each colour channel of the image. The package could be used by people to reduce the size of images, which could then be uploaded to social media platforms or other websites and applications.

Functions

• SeamCarve (class) Note : We plan to use some code provided for an assignment in DSCI 512 to implement seam carving. In particular, the functions energy and remove_vertical_seam/remove_horizontal_seam were provided.

  • seam_carve(img, height, width)
    • Description: This function will apply seam carving to the input image using Dynamic Programming to carve out lowest energy pixels depending on the input size given. The function will call other utility functions to calculate the minimum energy seams and remove them.
    • Input:
      • image (2d or 3d array)
      • desired_height (integer)
      • desired_width (integer)
    • Output:
      • compressed image (3d array, size desired_height x desired_width x 3)
  • energy(image)
    • Description: Computes the "energy" of an image, using the Laplacian of each colour channel and summing them up.
    • Input:
      • image: (a 3d array)
    • Output: energy (a 2d array)
  • find_vertical_seam(energy)/find_horizontal_seam(energy)
    • Description: This function uses dynamic programming to calculate the minimum energy vertical/horizontal seam.
    • Input:
      • energy (2d array)
    • Output:
      • seam (1d array, same as height/width in size)
  • remove_vertical_seam(image, seam)/remove_vertical_seam(image, seam)
    • Description: This function removes the calculated seam from the image.
    • Input:
      • image (3d array, size h x w x 3)
      • seam (1d array, same as height/width in size)
    • Output:
      • image (3d array, of size (h-1) x w x 3 / h x (w-1) x 3

• crop(image, height, width)

  • Description: This function will reduce the image to the specified size removing rows and columns of pixels from the borders.
  • Input:
    • image (3d array)
    • desired_height (integer)
    • desired_width (integer)
  • Output:
    • cropped image (3d array, size desired_height x desired_width x 3)

• compress(image, b = 4)

  • Description: This function compresses the image by reducing the number of bits for each channel based on user input.
  • Input:
    • image (3d array)
    • b (integer, range [1, 7] (number of bits used for each channel in the compressed image))
  • Output:
    • image (3d array, compressed to b bits)

• image_size(image)

  • Description: Calculates and returns the size of an image in bytes.
  • Input:
    • image (3d array)
  • Output:
    • size (integer, size in bytes)

There already are packages for image processing in R and Python:

• The magick package in R
• sckiti-image in Python

The existing packages are very comprehensive and provide many functions such as transformations, filters, file conversions and other advanced functions. Our packages focus specifically on image compression and reducing image size using Dynamic Programming and K-means Clustering.