Custom DeepSVDD for Image Classification

Custom Deep Support Vector Data Description(or DeepSVDD) for One-Class Classification(OCC) in Machine Learning (Ref. Deep OCC ICML 2018 paper).

This repository forked from lukasruff repository, I modified it to allow user to custom dataset.

Preparing Dataset

• Adding train list in train.csv in ./data/custom/ directory.
• Adding train list in test.csv in ./data/custom/ directory.
• Adding your image dataset to ./data/custom/ directory.

After prepared your dataset, let’s try it via my notebook.

Local Usage

Preparing Environment

$ python3 -m venv env
$ source env/bin/activate
$ pip3 install -r requirements.txt

Example

train.csv

path,label
normal/1.png,0
abnormal/2.png,1
...

test.csv

path,label
test/3.png,0
test/4.png,1
...

Train Data

$ sh run-custom.sh