neurophant/enimda-demo

ENIMDA usage demo

ENIMDA Demo

Demonstration for ENIMDA library

Setup environment and packages

python-3.6 -m venv .env

source .env/bin/activate

pip install -r requirements.txt

Run

python demo.py

Examples

Within the images folder in source folder you will find source images - with border (bordered) and without (clear)

In the detected folder there are the results of source images processing - each image has its borders outlined for visual demonstration

Accuracy

Using included examples each minimized to 300px and default threshold value (0.5), detection rate for bordered images is 81.2%, false detection rate for clear images is 6.9%

With parameter columns set to 0.05 (5%) detection becomes times faster and these rates are 74-78% and 6-8% respectively

Please be notified that using columns or frames parameters may lead to unstable result for border detection - it would depend on image nature and structure