Image Segmentation using U-Net, U-Net with skip connections and M-Net architectures

Overview

Brain-Image-Segmentation

Segmentation of brain tissues in MRI image has a number of applications in diagnosis, surgical planning, and treatment of brain abnormalities. However, it is a time-consuming task to be performed by medical experts. In addition to that, it is challenging due to intensity overlap between the different tissues caused by the intensity homogeneity and artifacts inherent to MRI. Tominimize this effect, it was proposed to apply histogram based preprocessing. The goal of this project was to develop a robust and automatic segmentation of the human brain.

To tackle the problem, I have used a Convolutional Neural Network (CNN) based approach. U-net is one of the most commonly used and best-performing architecture in medical image segmentation. This moodel consists of the 2-D implementation of the U-Net.The performance was evaluated using Dice Coefficient (DSC).

Dataset

This model was built for the following dataset: https://figshare.com/articles/brain_tumor_dataset/1512427

3064 T1-weighted contrast-inhanced images with three kinds of brain tumor are provided in the dataset.The three types of tumor are

1.Glioma 2.Pituitary Tumor 3.Meningioma

dataset

Model Architecture

The first half of the U-net is effectively a typical convolutional neural network like one would construct for an image classification task, with successive rounds of zero-padded ReLU-activated convolutions and ReLU-activated max-pooling layers. Instead of classification occurring at the "bottom" of the U, symmetrical upsampling and convolution layers are used to bring the pixel-wise prediction layer back to the original dimensions of the input image.

Here is the architecture for the 2D U-Net from the original publication mentioned earlier:

u-net-architecture

Here's an example of the correlation between my predictions in a single 2D plane:

Example 1: Example 2:
ground truth prediction

Libraries Used

The code has been tested with the following configuration

  • h5py == 2.10.0
  • keras == 2.3.1
  • scipy == 0.19.0
  • sckit-learn == 0.18.1
  • tensorflow == 2.2.0
  • tgpu == NVIDIA Tesla K80 (Google Colab)

The U-Net was based on this paper: https://arxiv.org/abs/1802.10508

Tips for improving model:

-The feature maps have been reduced so that the model will train using under 12GB of memory. If you have more memory to use, consider increasing the feature maps this will increase the complexity of the model (which will also increase its memory footprint but decrease its execution speed).

-If you choose a subset with larger tensors (e.g. liver or lung), it is recommended to add another maxpooling level (and corresponding upsampling) to the U-Net model. This will of course increase the memory requirements and decrease execution speed, but should give better results because it considers an additional recepetive field/spatial size.

-Consider different loss functions. The default loss function here is a binary_crossentropy. Different loss functions yield different loss curves and may result in better accuracy. However, you may need to adjust the learning_rate and number of epochs to train as you experiment with different loss functions.

-Try exceuting other U-Net architectures in the 2d/model folders.

Owner
Angad Bajwa
Angad Bajwa
Neural models of common sense. 🤖

Unicorn on Rainbow Neural models of common sense. This repository is for the paper: Unicorn on Rainbow: A Universal Commonsense Reasoning Model on a N

AI2 60 Jan 05, 2023
Code, Data and Demo for Paper: Controllable Generation from Pre-trained Language Models via Inverse Prompting

InversePrompting Paper: Controllable Generation from Pre-trained Language Models via Inverse Prompting Code: The code is provided in the "chinese_ip"

THUDM 101 Dec 16, 2022
Hierarchical Motion Encoder-Decoder Network for Trajectory Forecasting (HMNet)

Hierarchical Motion Encoder-Decoder Network for Trajectory Forecasting (HMNet) Our paper: https://arxiv.org/abs/2111.13324 We will release the complet

15 Oct 17, 2022
Bayesian inference for Permuton-induced Chinese Restaurant Process (NeurIPS2021).

Permuton-induced Chinese Restaurant Process Note: Currently only the Matlab version is available, but a Python version will be available soon! This is

NTT Communication Science Laboratories 3 Dec 17, 2022
An end-to-end library for editing and rendering motion of 3D characters with deep learning [SIGGRAPH 2020]

Deep-motion-editing This library provides fundamental and advanced functions to work with 3D character animation in deep learning with Pytorch. The co

1.2k Dec 29, 2022
YOLOV4运行在嵌入式设备上

在嵌入式设备上实现YOLO V4 tiny 在嵌入式设备上实现YOLO V4 tiny 目录结构 目录结构 |-- YOLO V4 tiny |-- .gitignore |-- LICENSE |-- README.md |-- test.txt |-- t

Liu-Wei 6 Sep 09, 2021
NLP made easy

GluonNLP: Your Choice of Deep Learning for NLP GluonNLP is a toolkit that helps you solve NLP problems. It provides easy-to-use tools that helps you l

Distributed (Deep) Machine Learning Community 2.5k Jan 04, 2023
Surrogate- and Invariance-Boosted Contrastive Learning (SIB-CL)

Surrogate- and Invariance-Boosted Contrastive Learning (SIB-CL) This repository contains all source code used to generate the results in the article "

Charlotte Loh 3 Jul 23, 2022
Keras + Hyperopt: A very simple wrapper for convenient hyperparameter optimization

This project is now archived. It's been fun working on it, but it's time for me to move on. Thank you for all the support and feedback over the last c

Max Pumperla 2.1k Jan 03, 2023
improvement of CLIP features over the traditional resnet features on the visual question answering, image captioning, navigation and visual entailment tasks.

CLIP-ViL In our paper "How Much Can CLIP Benefit Vision-and-Language Tasks?", we show the improvement of CLIP features over the traditional resnet fea

310 Dec 28, 2022
Official Pytorch implementation of "Learning Debiased Representation via Disentangled Feature Augmentation (Neurips 2021, Oral)"

Learning Debiased Representation via Disentangled Feature Augmentation (Neurips 2021, Oral): Official Project Webpage This repository provides the off

Kakao Enterprise Corp. 68 Dec 17, 2022
Kroomsa: A search engine for the curious

Kroomsa A search engine for the curious. It is a search algorithm designed to en

Wingify 7 Jun 20, 2022
SegNet model implemented using keras framework

keras-segnet Implementation of SegNet-like architecture using keras. Current version doesn't support index transferring proposed in SegNet article, so

185 Aug 30, 2022
PyTorch code for 'Efficient Single Image Super-Resolution Using Dual Path Connections with Multiple Scale Learning'

Efficient Single Image Super-Resolution Using Dual Path Connections with Multiple Scale Learning This repository is for EMSRDPN introduced in the foll

7 Feb 10, 2022
一个免费开源一键搭建的通用验证码识别平台,大部分常见的中英数验证码识别都没啥问题。

captcha_server 一个免费开源一键搭建的通用验证码识别平台,大部分常见的中英数验证码识别都没啥问题。 使用方法 python = 3.8 以上环境 pip install -r requirements.txt -i https://pypi.douban.com/simple gun

Sml2h3 189 Dec 02, 2022
This repository contains a CBIR system that uses swin transformer to extract image's feature.

Swin-transformer based CBIR This repository contains a CBIR(content-based image retrieval) system. Here we use Swin-transformer to extract query image

JsHou 12 Nov 17, 2022
2021搜狐校园文本匹配算法大赛 分比我们低的都是帅哥队

sohu_text_matching 2021搜狐校园文本匹配算法大赛Top2:分比我们低的都是帅哥队 本repo包含了本次大赛决赛环节提交的代码文件及答辩PPT,提交的模型文件可在百度网盘获取(链接:https://pan.baidu.com/s/1T9FtwiGFZhuC8qqwXKZSNA ,

hflserdaniel 43 Oct 01, 2022
LightLog is an open source deep learning based lightweight log analysis tool for log anomaly detection.

LightLog Introduction LightLog is an open source deep learning based lightweight log analysis tool for log anomaly detection. Function description [BG

25 Dec 17, 2022
Home for cuQuantum Python & NVIDIA cuQuantum SDK C++ samples

Welcome to the cuQuantum repository! This public repository contains two sets of files related to the NVIDIA cuQuantum SDK: samples: All C/C++ sample

NVIDIA Corporation 147 Dec 27, 2022
A spatial genome aligner for analyzing multiplexed DNA-FISH imaging data.

jie jie is a spatial genome aligner. This package parses true chromatin imaging signal from noise by aligning signals to a reference DNA polymer model

Bojing Jia 9 Sep 29, 2022