In this page you can find and download the code for LDNN, CHM and semi-supervised learning algorithms. If you use any of these packages please cite the following:

— "Regularization With Stochastic Transformations and Perturbations for Deep Semi-Supervised Learning" M. Sajjadi, M. Javanmardi, and T. Tasdizen, NIPS 2016

— "Image Segmentation with Cascaded Hierarchical Models and Logistic Disjunctive Normal Networks" M. Seyedhosseini, M. Sajjadi, and T. Tasdizen, ICCV 2013 [pdf, bibtex]

— "Disjunctive Normal Networks" M. Sajjadi, M. Seyedhosseini, and T. Tasdizen, Neurocomputing 2016

— The source code for both unsupervised loss functions are available here.

— This code is based on 'Spatially Sparse Convolutional Neural Networks' framework.

— Please refer to the 'README.md' file for instructions and installations.

— The source code and precompiled binaries for different platforms are available here.

— If you want to compile the mex files use “compile.m” script available in the package.

— We integrated the kmeans function of the berkeley vision group for clustering.

— ”LDNN_train.m” is the main function for training the classifier.

— “LDNN_predict.m” is the main function for testing the classifier.

— You can use “demo.m” to run some examples.

— The source code and precompiled binaries for 64-bit linux are available here.

— If you want to compile the mex files, use “compile.m” script available in the package.

— You should start with “TrainScript.m”.

— This code was written from scratch for the purpose of sharing.

— The results are not exactly the same as the reported numbers in the paper

(fortunately they are slightly better)