scholarly journals Medical Image Segmentation based on Fully Convolutional Network and Minimizing Energy Between Curves

TEM Journal ◽  
2020 ◽  
pp. 1348-1356
Author(s):  
Vo Thi Hong Tuyet ◽  
Nguyen Thanh Binh
Keyword(s):  
Image Segmentation ◽  
Medical Image ◽  
Medical Image Segmentation ◽  
Contour Detection ◽  
Gradient Vector Flow ◽  
Gradient Vector ◽  
High Quality ◽  
Convolutional Network ◽  
Fully Convolutional Network ◽  
Object Contour

Energy between curves of image has useful for object contour. The edge map is an important task for recognition. The shape that is found by linking between edges will clearly present the useful information of objects. The aim of medical image segmentation is the representation of a medical image into small pieces. In this process, feature extraction must adapt with edge map completely. This paper proposed a solution for medical image segmentation based on fully convolutional network with gradient vector flow snake in bandelet domain. Our approach depends on decomposition in bandelet domain and reconstruction in contour detection by fully convolutional network combining with gradient vector flow snake. To improve the accuracy of the feature's extraction processing, the proposed method detected the edge map in bandelet domain by using fully convolutional network. And its reconstructed objects contour by using gradient vector flow snake combined with the boundary condition. The results of the proposed method have the segmentation clearly with small details of medical images in high-quality and low-quality cases.

Interactive Medical Image Segmentation Method with only Positive Label Examples

2014 ◽  
Vol 989-994 ◽  
pp. 1088-1092
Author(s):  
Chen Guang Zhang ◽  
Yan Zhang ◽  
Xia Huan Zhang
Keyword(s):  
Image Segmentation ◽  
Medical Image ◽  
Medical Image Segmentation ◽  
Segmentation Method ◽  
High Quality ◽  
Neighbor Graph ◽  
Foreground Region ◽  
Training Examples ◽  
Positive Label ◽  
Negative Training

In this paper, a novel interactive medical image segmentation method called SMOPL is proposed. This method only needs marking some pixels on foreground region for segmentation. To do this, SMOPL characterize the inherent correlations among foreground and background pixels as Hilbert-Schmidt independence. By maximizing the independence and minimizing the smoothness of labels on instance neighbor graph simultaneously, SMOPL gets the sufficiently smooth confidences of both positive and negative classes in absence of negative training examples. Then a image segmentation can be obtained by assigning each pixel to the label for which the greatest confidence is calculated. Experiments on real-world medical images show that SMOPL is robust to get a high-quality segmentation with only positive label examples.

scholarly journals STAMP: Simultaneous Training and Model Pruning for Low Data Regimes in Medical Image Segmentation

2021 ◽  
Author(s):  
Nicola K Dinsdale ◽  
Mark Jenkinson ◽  
Ana IL Namburete
Keyword(s):  
Neural Networks ◽  
Image Segmentation ◽  
Medical Image ◽  
Medical Image Segmentation ◽  
High Quality ◽  
Vast Number ◽  
Large Numbers ◽  
Segmentation Algorithms ◽  
Simultaneous Training ◽  
Training And Pruning

Acquisition of high quality manual annotations is vital for the development of segmentation algorithms. However, to create them we require a substantial amount of expert time and knowledge. Large numbers of labels are required to train convolutional neural networks due to the vast number of parameters that must be learned in the optimisation process. Here, we develop the STAMP algorithm to allow the simultaneous training and pruning of a UNet architecture for medical image segmentation with targeted channelwise dropout to make the network robust to the pruning. We demonstrate the technique across segmentation tasks and imaging modalities. It is then shown that, through online pruning, we are able to train networks to have much higher performance than the equivalent standard UNet models while reducing their size by more than 85% in terms of parameters. This has the potential to allow networks to be directly trained on datasets where very low numbers of labels are available.

Sign in / Sign up

Export Citation Format

Share Document