Giter Site home page Giter Site logo

oodish's Introduction

OODISH

Pytorch implementation of Out-Of-Distribution Error Detection algorithms This repository is built for the paper 'Out of Distribution Detection in MedicalImaging via Convolutional Autoencoders' by Will Robinson.

This project is primarily divided into two IronPython notebooks which can be loaded using Jupyter or your interpreter of choice.

The first portion, autoencoder-pytorch_validate.ipynb loads a given dataset, constructs the CAE model with the given parameters, then runs the parameter sets until the specified number of epochs elapses or the early stopping criteria is met, or until memory runs out. It will save the best-performing model in the dataset folder and display some basic reconstructions so we can demonstrate it is working.

The second portion, load-anomaly-testing.ipynb will load the given model from the first notebook, perform some initial collection of distribution data for comparative purposes, then generate a specified number of test cases with a 50% chance of a given test case being adversarial. The model will then be used to reconstruct the image, the loss measured using MSE, SSIM and the encoding divergence measured using Mahalanobis distance, the ROC curve and accuracy scores will be generated and the loss comparison scatterplot and reconstruction histograms generated. Finally the model will output the first 10 images and their encodings and reconstructions, then all of the failures (FNs and FPs) for examination.

oodish's People

Contributors

aber-wgr avatar

Watchers

 avatar

Recommend Projects

  • React photo React

    A declarative, efficient, and flexible JavaScript library for building user interfaces.

  • Vue.js photo Vue.js

    ๐Ÿ–– Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.

  • Typescript photo Typescript

    TypeScript is a superset of JavaScript that compiles to clean JavaScript output.

  • TensorFlow photo TensorFlow

    An Open Source Machine Learning Framework for Everyone

  • Django photo Django

    The Web framework for perfectionists with deadlines.

  • D3 photo D3

    Bring data to life with SVG, Canvas and HTML. ๐Ÿ“Š๐Ÿ“ˆ๐ŸŽ‰

Recommend Topics

  • javascript

    JavaScript (JS) is a lightweight interpreted programming language with first-class functions.

  • web

    Some thing interesting about web. New door for the world.

  • server

    A server is a program made to process requests and deliver data to clients.

  • Machine learning

    Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.

  • Game

    Some thing interesting about game, make everyone happy.

Recommend Org

  • Facebook photo Facebook

    We are working to build community through open source technology. NB: members must have two-factor auth.

  • Microsoft photo Microsoft

    Open source projects and samples from Microsoft.

  • Google photo Google

    Google โค๏ธ Open Source for everyone.

  • D3 photo D3

    Data-Driven Documents codes.