scholarly journals The IoT and registration of MRI brain diagnosis based on genetic algorithm and convolutional neural network

Author(s):  
Ahmed Shihab Ahmed ◽  
Hussein Ali Salah

The technology <span>of the multimodal brain image registration is the key method for accurate and rapid diagnosis and treatment of brain diseases. For achieving high-resolution image registration, a fast sub pixel registration algorithm is used based on single-step discrete wavelet transform (DWT) combined with phase convolution neural network (CNN) to classify the registration of brain tumors. In this work apply the genetic algorithm and CNN clasifcation in registration of magnetic resonance imaging (MRI) image. This approach follows eight steps, reading the source of MRI brain image and loading the reference image, enhencment all MRI images by bilateral filter, transforming DWT image by applying the DWT2, evaluating (fitness function) each MRI image by using entropy, applying the genetic algorithm, by selecting the two images based on rollout wheel and crossover of the two images, the CNN classify the result of subtraction to normal or abnormal, “in the eighth one,” the Arduino and global system for mobile (GSM) 8080 are applied to send the message to patient. The proposed model is tested on MRI Medical City Hospital in Baghdad database consist 550 normal and 350 abnormal and split to 80% training and 20 testing, the proposed model result achieves the 98.8% </span>accuracy.

2017 ◽  
pp. 1437-1467
Author(s):  
Joydev Hazra ◽  
Aditi Roy Chowdhury ◽  
Paramartha Dutta

Registration of medical images like CT-MR, MR-MR etc. are challenging area for researchers. This chapter introduces a new cluster based registration technique with help of the supervised optimized neural network. Features are extracted from different cluster of an image obtained from clustering algorithms. To overcome the drawback regarding convergence rate of neural network, an optimized neural network is proposed in this chapter. The weights are optimized to increase the convergence rate as well as to avoid stuck in local minima. Different clustering algorithms are explored to minimize the clustering error of an image and extract features from suitable one. The supervised learning method applied to train the neural network. During this training process an optimization algorithm named Genetic Algorithm (GA) is used to update the weights of a neural network. To demonstrate the effectiveness of the proposed method, investigation is carried out on MR T1, T2 data sets. The proposed method shows convincing results in comparison with other existing techniques.


2014 ◽  
Vol 951 ◽  
pp. 274-277 ◽  
Author(s):  
Xu Sheng Gan ◽  
Can Yang ◽  
Hai Long Gao

To improve the optimization design of Radial Basis Function (RBF) neural network, a RBF neural network based on a hybrid Genetic Algorithm (GA) is proposed. First the hierarchical structure and adaptive crossover probability is introduced into the traditional GA algorithm for the improvement, and then the hybrid GA algorithm is used to optimize the structure and parameters of the network. The simulation indicates that the proposed model has a good modeling performance.


2011 ◽  
Vol 267 ◽  
pp. 19-24
Author(s):  
Hui Zhong Zhu ◽  
Yong Sheng Ding ◽  
Xiao Liang ◽  
Kuang Rong Hao ◽  
Hua Ping Wang

A novel neural network-based approach with immune genetic algorithm is proposed to conduct the optimizing design for the industrial filament manufacturing system. A new model is proposed in this paper to acquire better filament quality during such process. The proposed model was a combination of two components, namely, a traditional neural network which is used to simulate and an immune genetic algorithm-based part which is to improve the performance of the neural network component. Simulation results demonstrate that the proposed method can efficiently demonstrate the spinning process of filament and conduct the prediction of the filament quality with the production parameters as input data. Meanwhile, the proposed method enjoys faster speed and more precise accuracy, compared with traditional methods.


2021 ◽  
Author(s):  
Wei Li Zhang

From experiments, it is shown that the Co-occurrence matrix for one still MRI brain image does not provide enough information for segmentation. The 6D Co-occurrence image segmentation idea for 3D MRI image is modified and implemented in 2D MRI image segmentation. That idea is to take two or three images as input at the same time and then process them with 3D Co-occurrence matrix. With this kind of processing a lot of information was brought into the Co-occurrence matrix, which is enough to segment the images. To compare the result, some other segmentation ideas were tested in this project. From the results, it can be seen that the MRI image segmentation based on the Co-ocurrence texture analysis with two images or three images sampling is practical and the result satisfying. The segmentation is simulated in MATLAB. After the simulation, the segmentation is implemented in FPGA using VHDL. MODELSIM is used for FPGA functionality simulation. The result is close the MATLAB simulation. This makes it possible to implement the system with FPGA hardware.


2014 ◽  
Vol 631-632 ◽  
pp. 79-85 ◽  
Author(s):  
Feng Yu ◽  
Zhi Qing Wang ◽  
Xiao Zhong Xu

Aiming at the limitations of a single neural network for effective gas load forecasting, a combinational model based on wavelet BP neural network optimized by genetic algorithm is proposed. The problems that traditional BP algorithm converges slowly and easily falls into local minimum are overcame. The wavelet neural network strengthens the function approximation capacity of the network by combining the well time-frequency local feature of wavelet transform with the self-learning ability of neural network. And optimized by the real coded genetic algorithm, the network converges more quick than non-optimized one. This proposed model is applied to daily gas load forecasting for Shanghai and the simulation results indicate that this algorithm has excellent prediction effect.


2019 ◽  
Vol 8 (4) ◽  
pp. 10209-10218

Over last few decades, 3D reconstruction of medical images becomes advance technique in medical image processing. Reconstruction of 2D images of such data sets into 3D volumes, via registration of 2D sections had become a most interesting topic. In current years, MRI has been used for many medical analysis applications. The proposed system considered MRI images are taken from the same view, different times or acquired by different imaging modalities to increase the information. T1, T2 and PD MRI are taken as an input; T2 image is registered with a reference of T1 image using affine transformation, the registered T2 image is fused with T1 using DWT. The Fused T1T2 image is taken as reference image to register PD image using B-Spline transformation. DT-CWT technique is used to fuse the T1T2 image with registered PD image. The performance of the system shows that the proposed system gives more information by fusing T1T2PD images.


2021 ◽  
Author(s):  
Wei Li Zhang

From experiments, it is shown that the Co-occurrence matrix for one still MRI brain image does not provide enough information for segmentation. The 6D Co-occurrence image segmentation idea for 3D MRI image is modified and implemented in 2D MRI image segmentation. That idea is to take two or three images as input at the same time and then process them with 3D Co-occurrence matrix. With this kind of processing a lot of information was brought into the Co-occurrence matrix, which is enough to segment the images. To compare the result, some other segmentation ideas were tested in this project. From the results, it can be seen that the MRI image segmentation based on the Co-ocurrence texture analysis with two images or three images sampling is practical and the result satisfying. The segmentation is simulated in MATLAB. After the simulation, the segmentation is implemented in FPGA using VHDL. MODELSIM is used for FPGA functionality simulation. The result is close the MATLAB simulation. This makes it possible to implement the system with FPGA hardware.


Accurate and precise prediction of pricing of stock market is a very demanding task because of volatile, chaotic nature of time series data. Artificial Neural Networks played a major role for solving diversified problems for its robustness, strong capability for solving non linear problems and generalization ability. It is a popular choice for researchers for foretelling the financial time series data. In the article Pi Sigma Neural Network (PSNN) is developed for foretelling of stock market pricing in different time horizons. Pricing of stock market is predicted for one, fifteen and thirty days in advance. The parameters of the network are interpreted and optimized by Multiple Offspring Genetic Algorithm (MOGA). The motivation of this study is to achieve global optima with faster convergence rate. Bombay stock Exchange (BSE) data set is used for implementing the proposed model. Performance of the proposed model is evaluated using metrics like Mean Absolute Error (MAE), Root Mean Square Error (RMSE), Median Average Error (MedAE) . The results are compared with Pi Sigma Neural Network with Genetic Algorithm (PSNN-GA) and Pi Sigma Neural Network with Differential Evolution (PSNN-DE). It is concluded that the proposed model outperforms PSNN-GA and PSNN-DE models


Sign in / Sign up

Export Citation Format

Share Document