Image Processing Techniques for Diagnosing Rice Plant Disease: A Survey

Brain tumor is a fatal disease, caused by the growth of abnormal cells in the brain tissues. Therefore, early and accurate detection of this disease can save patient’s life. This paper proposes a novel framework for the detection of brain tumor using magnetic resonance (MR) images. The framework is based on the fully convolutional neural network (FCNN) and transfer learning techniques. The proposed framework has five stages which are preprocessing, skull stripping, CNN-based tumor segmentation, postprocessing, and transfer learning-based brain tumor binary classification. In preprocessing, the MR images are filtered to eliminate the noise and are improve the contrast. For segmentation of brain tumor images, the proposed CNN architecture is used, and for postprocessing, the global threshold technique is utilized to eliminate small nontumor regions that enhanced segmentation results. In classification, GoogleNet model is employed on three publicly available datasets. The experimental results depict that the proposed method is achieved average accuracies of 96.50%, 97.50%, and 98% for segmentation and 96.49%, 97.31%, and 98.79% for classification of brain tumor on BRATS2018, BRATS2019, and BRATS2020 datasets, respectively. The outcomes demonstrate that the proposed framework is effective and efficient that attained high performance on BRATS2020 dataset than the other two datasets. According to the experimentation results, the proposed framework outperforms other recent studies in the literature. In addition, this research will uphold doctors and clinicians for automatic diagnosis of brain tumor disease.

/pdf/automated-detection-of-brain-tumor-through-magnetic-1he72b6pjr.pdf

Automated Detection of Brain Tumor through Magnetic Resonance Images Using Convolutional Neural Network

Brain tumor segmentation, an essential but challenging task, has long attracted much attention from the medical imaging community. Recently, successful applications of sparse coding and dictionary learning has emerged in various vision problems including image segmentation. In this paper, a superpixel-based framework for automated brain tumor segmentation is introduced. The kernel trick is adopted in dictionary learning to transform superpixel-level features to a high-dimensional feature space where their nonlinear similarities are considered to generate discriminative sparse codes. A graph is constructed from the approximation errors given by dictionaries modeling different brain tumor structures so that superpixels belonging to particular tumor regions can be efficiently identified. The proposed framework is evaluated on brain magnetic resonance images of high-grade glioma (HGG) patients provided by the multi-modal Brain Tumor Segmentation (BRATS) Benchmark. Results show that the proposed framework achieves competitive performance when compared with the state-of-the-art methods.

Automated brain tumor segmentation using kernel dictionary learning and superpixel-level features

With the development of healthcare technologies, the elderly population has grown and therefore populating ageing has emerged as a social issue. It is a cause of rise in patients with geriatric disorders, among which dementia is very fatal to the elderly’s activities of daily living. In the studies on dementia risk prediction, a method using deep learning was proposed. It requires a lot of image data and much time to learn. Therefore, this study proposes a prediction model of dementia risk based on XGBoost using derived variable extraction from numericalized dementia data and hyper-parameters optimization. The proposed method extracts variable importance from typical independent variables with the use of gradient boosting and then generates derived variables. The generated derived variables are applied to variable importance analysis and thereby a Top-N group is created. Then, for achieving optimal performance in line with the data characteristics of each Top-N group, hyper-parameter tuning is conducted. With the optimized groups, XGBoost model based performance is evaluated. In addition, for the performance evaluation of the proposed model, goodness-of-fit for machine learning classification models is evaluated. According to the Top-N group performance evaluation with different numbers of derived variables, Top-20 model showed the best performance, and the optimized hyper-parameter values were eta = 0.10, gamma = 0, max_depth = 4, and min_child_weight = 1. As a result, the accuracy of the XGBoost model proposed in this study was 85.61%, and its F1-score was 79.28%. When the proposed model is compared with Decision Tree, Random Forest, SVM, and k-NN models, it has the best performance.

/pdf/prediction-model-of-dementia-risk-based-on-xgboost-using-44zgmjrlra.pdf

Prediction Model of Dementia Risk Based on XGBoost Using Derived Variable Extraction and Hyper Parameter Optimization

A Brain Computer Interface (BCI) is at its early stage in their development, but the new significant improvement needs to be taken to help bring back the quality of moving freely in a person with a disability The BCI is a captivating application which could be used in various fields like Gaming, Education, Industrial and Medical areas In this paper, EEG-based Brain controlled Wheelchair has been developed using BCI with the help of Neuro sky technology The Event-Related Potential (ERP) offered the Neurosignal information with the help of P300 component BCI system consists of a Biosensor to capture the EEG/EMG signals The signals will be processed by the ThinkGear module in MATLAB The Level Analyzer Technique is performed on all the training signals and Alpha and Beta waves are extracted for controlling the wheelchair The command signals are transmitted to the Microprocessor via RF medium The robotic module designed consists of ARM 7 Microprocessor coupled with DC motor to perform the command The Eye blinking strength and attention level were used to control the direction of the wheelchair The wireless BCI system could allow the paralyzed people to control their wheelchair without any complexity, provided it is more enhanced, portable and wearable

Electroencephalogram-based Brain Controlled Robotic Wheelchair

A neural network based deep learning approach for efficient segmentation of brain tumor medical image data

Robots are playing a vital role in today's industrial automation and monitoring system. As technology developed these robots have increased their applications and functionality. Working robots will cooperate to the people makes the work more effortless and uncomplicated. This paper provides 4 different gestures for controlling the robots, i.e., forward, backward, left, right. For cutting weeds a gripper concept using buttons is anticipated. These movements are given by the user using MEMS accelerometer. The MEMS accelerometer will be set to the hand. Whenever the hand moves in some direction, the mechanical movement of the hand will be recognized by MEMS. MEMS translate this mechanical hand movement into equivalent electrical signals and send it to the PIC microcontroller. The PIC microcontroller at the transmitter side sends control signals to the receiver side through RF TRANSCEIVER. The controller at the receiver area receives these signals and gives direction to the robot. This type of robot is used in the crop filed to cut the weeds as per the user command.

Gesture Controlled Robot using MEMS Accelerometer for Eradication of Weeds

In data mining, the clustering is one of the important process for categorizing the elements into groups whose associated members are similar in their features. In this paper, the plant leaves are grouped based on the colors in the leaves. Totally, three categories are specified to represent the leaf with more green, leaf with yellowish shades and leaf with reddish shades. The task is performed using image processing. The leaf images are processed in the sequence such as image pre-processing, segmentation, feature extraction and clustering. Pre-processing is done to denoize, enhance and background color fixing for betterment of result. Then, the color-based segmentation is done on the pre-processed image for generating the sub-images by clustering the pixels based on the colors. Next, the basic features such as entropy, mean and stadard deviation are extracted from each sub-images. The extracted features are used for clustering the images based on the colors. The image clustering is done by the Neural Network architecture, self-organizing map (SOM) and K-means algorithm. They are evaluated with various distance measuring functions. Finally, the city-block in both method produced the clusters with same size. This cluster set can be used as a training set for the leaf classification in future.

Unsupervised machine learning for clustering the infected leaves based on the leaf-colours

Traffic safety in the area of Intelligent Transportation System can be improved by predicting driver behavior automatically. Aggressive driving is significant indicator for collision. Several deep learning algorithms are developed to predict the behavior of driver and existing algorithm’s hyper parameters are not optimized to get efficient and accurate solution in predicting driver’s behavior. Hyper parameter optimization of LSTM based model has been proposed to predict the driver behavior accurately by determining optimal hyper parameters which includes window size, learning rate, Number of hidden layers and number of hidden units. Bayesian optimization is built to optimize the hyper parameters. 97.02% of accuracy has been achieved for the proposed model. The result reveals that the proposed model outperforms the existing algorithms and hyper parameters have been optimized. The hyper parameter optimization can improve the accuracy of the LSTM based driver’s aggressive behavior prediction model. The proposed model minimizes False alarm rate and missed alarm rate dramatically compared to traditional algorithms. The proposed model can be implemented in collision avoidance system which helps to present warning to aggressive drivers.

Hyperparameter optimization of LSTM based Driver’s Aggressive Behavior Prediction Model

The Almost 1.3 million casualties are reported round a calendar year due to road accidents. Advanced collision avoidance systems play major role in predicting the collision risk to avoid accidents. The existing deep learning algorithms are unable to predict the crash risk efficiently. In the existing system, Long Short Term Memory algorithm is used to predict the crash risk where weights are not optimized. The objective is to predict the rear end collision risk with optimized weight by combining Long Short Term Memory(LSTM) with Levenberg–Marquardt (LM) algorithms. The proposed algorithm predicts the collision risk considering vehicle, driver related factors, and temporal dependencies. Next Generation Simulation Project (NGSIM) dataset is used to evaluate the proposed model. The performance of the proposed system is compared with the performance of Long Short Term Memory and Back Propagation Neural Network. 95.6% of accuracy is achieved by LM-LSTM based Time series Deep Network Model. The prediction accuracy has been improved considerably than the existing algorithms. There is the drastic improvement in minimization of false alarm and missed alarm rate. The main advantage of the proposed system is that it will present warning at the time of high collision risk and it helps drivers to prevent from accident.

K. Ashok Kumar

Papers

A neural network based deep learning approach for efficient segmentation of brain tumor medical image data

Gesture Controlled Robot using MEMS Accelerometer for Eradication of Weeds

Unsupervised machine learning for clustering the infected leaves based on the leaf-colours

Hyperparameter optimization of LSTM based Driver’s Aggressive Behavior Prediction Model

Levenberg–Marquardt –LSTM based Efficient Rear-end Crash Risk Prediction System Optimization