falreis/segmentation-eval

Evaluates deep neural networks for oversegmentation and semantic segmentation

Combining convolutional side-outputs for image segmentation and border detection

Source Code

To configure the virtual environment, import the Anaconda i2dl.yml .yml file, as instructed below. The evaluation environment only runs with Python 3.

conda env create -f i2dl.yml

The step-by-step approach to running the neural network already trained, training for the Kitti and BSDS500 databases, and performance evaluation is available in the source code tutorial.

KITTI Performance Evaluation

To configure the virtual environment, import the Anaconda kitti.yml file, as described below. The evaluation environment only runs with Python 2.

conda env create -f kitty.yml

Papers

It will available soon.