An image compressor in Python. This program uses the unsupervised K-means algorithm. The dockerized application is then deployed on a raspberry pi 4, an ege device.
In this project, I have shown that a normal image takes 4 seconds to upload while a compressed image takes 2 seconds. This shows that processing an image at edge takes less time thereby implementing low latency.
The container does the following things:
- Take an image file.
- Upload it to amazon S3.
- Compresses the image using unsupervised K-means.
- Upload the compressed image to amazon S3.
- To start the container, execute:
With docker-compose
$ docker-compose up --build
Without docker-compose
Unzip the project folder
Open the project folder
In linux
cd PythonImageCompressor
Build the docker image
docker build -t imageproc .
To run the container
docker run imageproc
Output
