img

img is a simple image manipulation tool. It aims to provide a cli tool allowing for quick image manipulation as well as a rust library for applying image filters.

Features

img is a very young tool so support for different types of filters is limited to:

• blur (available algorithms: mean blur, gaussian blur)
• grayscale filter
• sepia filter
• gamma correction filter
• canny edge detection
• kuwahara filter

it also allows for simple image manipulation:

• crop
• resize (using nearest algorithm)

in terms of I/O, reading and writing is implemented for:

• png files:
  • reading is implemented for grayscale, grayscale with alpha, rgb and rgba images with pixel depth of eight.
  • writing always results in an rgba eight bit depth image

Cli installation

clone the repository using git

git clone git@github.com:sobczal2/img.git

within repository directory invoke

cargo install --path crates/img-cli

to install the tool.

By default parallel feature is enabled which allows each operation to specify -t <thread_count> parameter where thread_count is either a positive number or auto (which is the default and uses number equal to logical cores available).

Cli Usage

Print Help

img --help

Blur:

img blur -i input.png -o output.png [algorithm] -r <radius> -s <sigma>

• algorithm - either "mean" or "gaussian"
• radius - non-negative integer describing range of the blur kernel. Passing value of 0 results in 1x1 kernel, value 1 - 3x3, etc.
• sigma - sigma value used in gaussian blur algorithm

Grayscale

img grayscale -i input.png -o output.png

Sepia

img sepia -i input.png -o output.png

Gamma Correction

img gamma-correction -i input.png -o output.png -g <gamma>

Crop

img crop -i input.png -o output.png -s [width]x[height]+[offset_x]x[offset_y]

• width - width of the target image in pixels
• height - height of the target image in pixels
• offset_x - horizontal offset of the target image in pixels (counting from left edge of the original image)
• offset_y - vertical offset of the target image in pixels (counting from top edge of the original image)

for example

img crop -i input.png -o output.png -s 100x200+10x20

results in an image 100x200 starting 10 pixels from the left edge and 20 pixels from the top edge

Resize

img resize -i input.png -o output.png -s [width]x[height]

• width - width of the target image in pixels
• height - height of the target image in pixels

Canny

img canny -i input.png -o output.png

Canny

img canny -i input.png -o output.png

Kuwahara

img kuwahara -i input.png -o output.png

Library usage

Image struct is the main struct holding image data. It holds RGBA images where each pixel value ranges from 0 to 255.

All filters use lens api (see Lens trait. This is a main way of interacting with image, each Lens transforms each point of an Image (or a different 2d representation). This api is lazy, inspired by Iterator so it does not perform any expensive calculations unless Lens::look method is called.

From there, you can follow code documentation.

Parallelism

This project can be compiled with "parallel" feature flag which adds corresponding functions utilizing parallelism, most important being FromLensPar::from_lens_par method.