Arraymancer Vision (WIP)

Simple library for image loading, preprocessing and visualization for working with arraymancer.

Features

• Loading image into tensors
• Simple image transformations like flipping, rotation, scaling
• Saving images
• Image convolution filters like sharpen, edges
• Visualization of images using visdom

Quick Start

Installation

Install using nimble package manager:

nimble install arraymancer_vision

For visualizing you have to install visdom and run it:

pip install visdom
python -m visdom.server

Then go to http://localhost:8097

Usage example

import arraymancer_vision

# Load image from file into a CxHxW Tensor[uint8]
var origimage = load("examples/lena.png")

# Do some preprocessing
var image = origimage.center_crop(128, 128)
image = image.hflip()
image = image.rot90(1)
image = image.filter_sharpen()
image = image.scale(512, 512, ScaleBilinear)

# Visualize it using visdom
let vis = newVisdomClient()
vis.image(origimage)
vis.image(image)

# Save it to a file
image.save("examples/preprocessed_lena.png")

This quickstart example is inside examples directory, you can run it by cloning the repo and running with nim c -r examples/quickstart.nim

You can visualize all predefined filters having visdom running and then running the filters example with nim c -r examples/visualize_filters.nim

API

Documentation of the completely available API is here

Details

The library operates all images as Tensor[uint8] with dimensions CxHxW, where C is in RGBA colorspace, note that other image libraries usually operates with images in HxWxC format, so remember this when using. This design choice is to optimize and facilitate operation on images in deep learning tasks.

TODO

• Simple drawing routines
• Colorspace conversions