Audio/Visual Library in Scala
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Failed to load latest commit information.

Sclaav, An Audio/Visual Library In Scala

A library and application for manipulating and creating images and audio.


Functionality which presently works well and default parameters are tuned to work well for most cases.

What Where Test
Image mosaic DoMosaic DoesMosaicTest
Image similarity ImageMatchers ImageMatchersTest
Docker image On Dockerhub

Experimental Functionality:

Functionality which may or may not work well yet, where default parameters are a WIP, etc.

What Where Test
Genetic Algorithm for sub-image matching SimpleCompleteGeneticAssembler SimpleCompleteGeneticAssemblerTest
Image mosaic using non-uniform quadrilateral grid QuadrilateralAssembler QuadrilateralAssemblerTest


$ ./sbt clean assembly

will create the far jar: bin/Sclaav.jar

Running Tests

$ ./sbt clean test

Using via CLI

This is the preferred method for using this tool.


Original Mosaic
  • Given a target image (or set of target images) and a set of sample images,
    • Discretize target image in grid by rows and cols.
    • For each sub-image, find the sample image which is the closest match; by default we use min-2-norm distance between ARGB average vectors.
    • Assemble each matched sample image onto the original to create the mosaic.
    • Eyeballs be amazed.

Run with:

$ java -jar bin/Sclaav.jar \ # invokes jar
  --in file:///input-files-dir \     # input images directory
  --out file:///output-dir \         # output directory
  --mode permute \                   # permute or single mode
  --rows 64 \                        # how many rows to split target
  --cols 64 \                        # how many cols to split target
  --samples 300 \                    # how many samples to use from input dir
  --filters true \                   # apply and use filtered images
  --target file:///photo.jpg         # if mode is single, specify the target image to use


Given a target image, print out the images which best matches based on a distance metric and similarity measure (Euclidean ARGB for now).

Run with:

$ java -jar bin/Sclaav.jar \ 
  --in file:///input-files-dir \     
  --out file:///output-dir \         
  --mode similarity \            
  --samples 300 \            
  --target file:///photo.jpg         


This comes with a Docker image! Build with:

./sbt docker

Or pull the image from Docker Hub:

docker pull ogeagla/sclaav

The container contains:

  • The assembly jar in /opt/sclaav/Sclaav.jar
  • DCRAW installation. Test with /opt/dcraw/scripts/

To use the container as a DCRAW execution environment, first start the container, then run a shell into it.

Run the container with:

docker run oct/sclaav:latest

Check that it is running with:

docker ps

Find the container ID and then run:

docker exec -it <container-id> bash

You will then have a prompt inside the running container. Run the test script with:

sh /opt/dcraw/scripts/

Which should create a thumbnail jpeg in /opt/dcraw/samples.



Apache 2.0