Classification of Citrus Plant Diseases Using Deep Transfer Learning

In recent years,the field of deep learning has played an important role towards automatic detection and classification of diseases in vegetables and fruits.This in turn has helped in improving the quality and production of vegetables and fruits.Citrus fruits arewell known for their taste and nutritional values.They are one of the natural and well known sources of vitamin C and planted worldwide.There are several diseases which severely affect the quality and yield of citrus fruits.In this paper,a new deep learning based technique is proposed for citrus disease classification.Two different pre-trained deep learning models have been used in this work.To increase the size of the citrus dataset used in this paper,image augmentation techniques are used.Moreover,to improve the visual quality of images,hybrid contrast stretching has been adopted.In addition,transfer learning is used to retrain the pre-trainedmodels and the feature set is enriched by using feature fusion.The fused feature set is optimized using a meta-heuristic algorithm,the Whale Optimization Algorithm(WOA).The selected features are used for the classification of six different diseases of citrus plants.The proposed technique attains a classification accuracy of 95.7%with superior results when compared with recent techniques.

Multi-Layered Deep Learning Features Fusion for Human Action Recognition

Human Action Recognition(HAR)is an active research topic in machine learning for the last few decades.Visual surveillance,robotics,and pedestrian detection are the main applications for action recognition.Computer vision researchers have introduced many HAR techniques,but they still face challenges such as redundant features and the cost of computing.In this article,we proposed a new method for the use of deep learning for HAR.In the proposed method,video frames are initially pre-processed using a global contrast approach and later used to train a deep learning model using domain transfer learning.The Resnet-50 Pre-Trained Model is used as a deep learning model in this work.Features are extracted from two layers:Global Average Pool(GAP)and Fully Connected(FC).The features of both layers are fused by the Canonical Correlation Analysis(CCA).Then features are selected using the Shanon Entropy-based threshold function.The selected features are finally passed to multiple classifiers for final classification.Experiments are conducted on five publicly available datasets as IXMAS,UCF Sports,YouTube,UT-Interaction,and KTH.The accuracy of these data sets was 89.6%,99.7%,100%,96.7%and 96.6%,respectively.Comparison with existing techniques has shown that the proposed method provides improved accuracy for HAR.Also,the proposed method is computationally fast based on the time of execution.

Dynamic Hand Gesture Recognition Using 3D-CNN and LSTM Networks

Recognition of dynamic hand gestures in real-time is a difficult task because the system can never know when or from where the gesture starts and ends in a video stream.Many researchers have been working on visionbased gesture recognition due to its various applications.This paper proposes a deep learning architecture based on the combination of a 3D Convolutional Neural Network(3D-CNN)and a Long Short-Term Memory(LSTM)network.The proposed architecture extracts spatial-temporal information from video sequences input while avoiding extensive computation.The 3D-CNN is used for the extraction of spectral and spatial features which are then given to the LSTM network through which classification is carried out.The proposed model is a light-weight architecture with only 3.7 million training parameters.The model has been evaluated on 15 classes from the 20BN-jester dataset available publicly.The model was trained on 2000 video-clips per class which were separated into 80%training and 20%validation sets.An accuracy of 99%and 97%was achieved on training and testing data,respectively.We further show that the combination of 3D-CNN with LSTM gives superior results as compared to MobileNetv2+LSTM.

An Integrated Deep Learning Framework for Fruits Diseases Classification

:Agriculture has been an important research area in the field of image processing for the last five years.Diseases affect the quality and quantity of fruits,thereby disrupting the economy of a country.Many computerized techniques have been introduced for detecting and recognizing fruit diseases.However,some issues remain to be addressed,such as irrelevant features and the dimensionality of feature vectors,which increase the computational time of the system.Herein,we propose an integrated deep learning framework for classifying fruit diseases.We consider seven types of fruits,i.e.,apple,cherry,blueberry,grapes,peach,citrus,and strawberry.The proposed method comprises several important steps.Initially,data increase is applied,and then two different types of features are extracted.In the first feature type,texture and color features,i.e.,classical features,are extracted.In the second type,deep learning characteristics are extracted using a pretrained model.The pretrained model is reused through transfer learning.Subsequently,both types of features are merged using the maximum mean value of the serial approach.Next,the resulting fused vector is optimized using a harmonic threshold-based genetic algorithm.Finally,the selected features are classified using multiple classifiers.An evaluation is performed on the PlantVillage dataset,and an accuracy of 99%is achieved.A comparison with recent techniques indicate the superiority of the proposed method.

Recognition and Tracking of Objects in a Clustered Remote Scene Environment

Object recognition and tracking are two of the most dynamic research sub-areas that belong to the field of Computer Vision.Computer vision is one of the most active research fields that lies at the intersection of deep learning and machine vision.This paper presents an efficient ensemble algorithm for the recognition and tracking of fixed shapemoving objects while accommodating the shift and scale invariances that the object may encounter.The first part uses the Maximum Average Correlation Height(MACH)filter for object recognition and determines the bounding box coordinates.In case the correlation based MACH filter fails,the algorithms switches to a much reliable but computationally complex feature based object recognition technique i.e.,affine scale invariant feature transform(ASIFT).ASIFT is used to accommodate object shift and scale object variations.ASIFT extracts certain features from the object of interest,providing invariance in up to six affine parameters,namely translation(two parameters),zoom,rotation and two camera axis orientations.However,in this paper,only the shift and scale invariances are used.The second part of the algorithm demonstrates the use of particle filters based Approximate Proximal Gradient(APG)technique to periodically update the coordinates of the object encapsulated in the bounding box.At the end,a comparison of the proposed algorithm with other stateof-the-art tracking algorithms has been presented,which demonstrates the effectiveness of the proposed algorithm with respect to the minimization of tracking errors.

Brain Tumor Detection and Classification Using PSO and Convolutional Neural Network

Tumor detection has been an active research topic in recent years due to the high mortality rate.Computer vision(CV)and image processing techniques have recently become popular for detecting tumors inMRI images.The automated detection process is simpler and takes less time than manual processing.In addition,the difference in the expanding shape of brain tumor tissues complicates and complicates tumor detection for clinicians.We proposed a newframework for tumor detection aswell as tumor classification into relevant categories in this paper.For tumor segmentation,the proposed framework employs the Particle Swarm Optimization(PSO)algorithm,and for classification,the convolutional neural network(CNN)algorithm.Popular preprocessing techniques such as noise removal,image sharpening,and skull stripping are used at the start of the segmentation process.Then,PSO-based segmentation is applied.In the classification step,two pre-trained CNN models,alexnet and inception-V3,are used and trained using transfer learning.Using a serial approach,features are extracted from both trained models and fused features for final classification.For classification,a variety of machine learning classifiers are used.Average dice values on datasets BRATS-2018 and BRATS-2017 are 98.11 percent and 98.25 percent,respectively,whereas average jaccard values are 96.30 percent and 96.57%(Segmentation Results).The results were extended on the same datasets for classification and achieved 99.0%accuracy,sensitivity of 0.99,specificity of 0.99,and precision of 0.99.Finally,the proposed method is compared to state-of-the-art existingmethods and outperforms them.

Automatic Heart Disease Detection by Classification of Ventricular Arrhythmias on ECG Using Machine Learning

This paper focuses on detecting diseased signals and arrhythmias classification into two classes:ventricular tachycardia and premature ventricular contraction.The sole purpose of the signal detection is used to determine if a signal has been collected from a healthy or sick person.The proposed research approach presents a mathematical model for the signal detector based on calculating the instantaneous frequency(IF).Once a signal taken from a patient is detected,then the classifier takes that signal as input and classifies the target disease by predicting the class label.While applying the classifier,templates are designed separately for ventricular tachycardia and premature ventricular contraction.Similarities of a given signal with both the templates are computed in the spectral domain.The empirical analysis reveals precisions for the detector and the applied classifier are 100%and 77.27%,respectively.Moreover,instantaneous frequency analysis provides a benchmark that IF of a normal signal ranges from 0.8 to 1.1 Hz whereas IF range for ventricular tachycardia and premature ventricular contraction is 0.08–0.6 Hz.This indicates a serious loss of high-frequency contents in the spectrum,implying that the heart’s overall activity is slowed down.This study may help medical practitioners in detecting the heart disease type based on signal analysis.

An Ensemble of Optimal Deep Learning Features for Brain Tumor Classification

Owing to technological developments,Medical image analysis has received considerable attention in the rapid detection and classification of diseases.The brain is an essential organ in humans.Brain tumors cause loss of memory,vision,and name.In 2020,approximately 18,020 deaths occurred due to brain tumors.These cases can be minimized if a brain tumor is diagnosed at a very early stage.Computer vision researchers have introduced several techniques for brain tumor detection and classification.However,owing to many factors,this is still a challenging task.These challenges relate to the tumor size,the shape of a tumor,location of the tumor,selection of important features,among others.In this study,we proposed a framework for multimodal brain tumor classification using an ensemble of optimal deep learning features.In the proposed framework,initially,a database is normalized in the form of high-grade glioma(HGG)and low-grade glioma(LGG)patients and then two pre-trained deep learning models(ResNet50 and Densenet201)are chosen.The deep learning models were modified and trained using transfer learning.Subsequently,the enhanced ant colony optimization algorithm is proposed for best feature selection from both deep models.The selected features are fused using a serial-based approach and classified using a cubic support vector machine.The experimental process was conducted on the BraTs2019 dataset and achieved accuracies of 87.8%and 84.6%for HGG and LGG,respectively.The comparison is performed using several classification methods,and it shows the significance of our proposed technique.

Enhanced Accuracy for Motor Imagery Detection Using Deep Learning for BCI

Brain-Computer Interface(BCI)is a system that provides a link between the brain of humans and the hardware directly.The recorded brain data is converted directly to the machine that can be used to control external devices.There are four major components of the BCI system:acquiring signals,preprocessing of acquired signals,features extraction,and classification.In traditional machine learning algorithms,the accuracy is insignificant and not up to the mark for the classification of multi-class motor imagery data.The major reason for this is,features are selected manually,and we are not able to get those features that give higher accuracy results.In this study,motor imagery(MI)signals have been classified using different deep learning algorithms.We have explored two different methods:Artificial Neural Network(ANN)and Long Short-Term Memory(LSTM).We test the classification accuracy on two datasets:BCI competition III-dataset IIIa and BCI competition IV-dataset IIa.The outcome proved that deep learning algorithms provide greater accuracy results than traditional machine learning algorithms.Amongst the deep learning classifiers,LSTM outperforms the ANN and gives higher classification accuracy of 96.2%.

Segmentation and Classification of Stomach Abnormalities Using Deep Learning

An automated system is proposed for the detection and classification of GI abnormalities.The proposed method operates under two pipeline procedures:(a)segmentation of the bleeding infection region and(b)classification of GI abnormalities by deep learning.The first bleeding region is segmented using a hybrid approach.The threshold is applied to each channel extracted from the original RGB image.Later,all channels are merged through mutual information and pixel-based techniques.As a result,the image is segmented.Texture and deep learning features are extracted in the proposed classification task.The transfer learning(TL)approach is used for the extraction of deep features.The Local Binary Pattern(LBP)method is used for texture features.Later,an entropy-based feature selection approach is implemented to select the best features of both deep learning and texture vectors.The selected optimal features are combined with a serial-based technique and the resulting vector is fed to the Ensemble Learning Classifier.The experimental process is evaluated on the basis of two datasets:Private and KVASIR.The accuracy achieved is 99.8 per cent for the private data set and 86.4 percent for the KVASIR data set.It can be confirmed that the proposed method is effective in detecting and classifying GI abnormalities and exceeds other methods of comparison.

Statistical Histogram Decision Based Contrast Categorization of Skin Lesion Datasets Dermoscopic Images

Most of the melanoma cases of skin cancer are the life-threatening form of cancer.It is prevalent among the Caucasian group of people due to their light skin tone.Melanoma is the second most common cancer that hits the age group of 15–29 years.The high number of cases has increased the importance of automated systems for diagnosing.The diagnosis should be fast and accurate for the early treatment of melanoma.It should remove the need for biopsies and provide stable diagnostic results.Automation requires large quantities of images.Skin lesion datasets contain various kinds of dermoscopic images for the detection of melanoma.Three publicly available benchmark skin lesion datasets,ISIC 2017,ISBI 2016,and PH2,are used for the experiments.Currently,the ISIC archive and PH2 are the most challenging and demanding dermoscopic datasets.These datasets’pre-analysis is necessary to overcome contrast variations,under or over segmented images boundary extraction,and accurate skin lesion classification.In this paper,we proposed the statistical histogram-based method for the pre-categorization of skin lesion datasets.The image histogram properties are utilized to check the image contrast variations and categorized these images into high and low contrast images.The two performance measures,processing time and efficiency,are computed for evaluation of the proposed method.Our results showed that the proposed methodology improves the pre-processing efficiency of 77%of ISIC 2017,67%of ISBI 2016,and 92.5%of PH2 datasets.

Anomalous Situations Recognition in Surveillance Images Using Deep Learning

Anomalous situations in surveillance videos or images that may result in security issues,such as disasters,accidents,crime,violence,or terrorism,can be identified through video anomaly detection.However,differentiat-ing anomalous situations from normal can be challenging due to variations in human activity in complex environments such as train stations,busy sporting fields,airports,shopping areas,military bases,care centers,etc.Deep learning models’learning capability is leveraged to identify abnormal situations with improved accuracy.This work proposes a deep learning architecture called Anomalous Situation Recognition Network(ASRNet)for deep feature extraction to improve the detection accuracy of various anomalous image situations.The proposed framework has five steps.In the first step,pretraining of the proposed architecture is performed on the CIFAR-100 dataset.In the second step,the proposed pre-trained model and Inception V3 architecture are used for feature extraction by utilizing the suspicious activity recognition dataset.In the third step,serial feature fusion is performed,and then the Dragonfly algorithm is utilized for feature optimization in the fourth step.Finally,using optimized features,various Support Vector Machine(SVM)and K-Nearest Neighbor(KNN)based classification models are utilized to detect anomalous situations.The proposed framework is validated on the suspicious activity dataset by varying the number of optimized features from 100 to 1000.The results show that the proposed method is effective in detecting anomalous situations and achieves the highest accuracy of 99.24%using cubic SVM.

HybridHR-Net:Action Recognition in Video Sequences Using Optimal Deep Learning Fusion Assisted Framework

The combination of spatiotemporal videos and essential features can improve the performance of human action recognition(HAR);however,the individual type of features usually degrades the performance due to similar actions and complex backgrounds.The deep convolutional neural network has improved performance in recent years for several computer vision applications due to its spatial information.This article proposes a new framework called for video surveillance human action recognition dubbed HybridHR-Net.On a few selected datasets,deep transfer learning is used to pre-trained the EfficientNet-b0 deep learning model.Bayesian optimization is employed for the tuning of hyperparameters of the fine-tuned deep model.Instead of fully connected layer features,we considered the average pooling layer features and performed two feature selection techniques-an improved artificial bee colony and an entropy-based approach.Using a serial nature technique,the features that were selected are combined into a single vector,and then the results are categorized by machine learning classifiers.Five publically accessible datasets have been utilized for the experimental approach and obtained notable accuracy of 97%,98.7%,100%,99.7%,and 96.8%,respectively.Additionally,a comparison of the proposed framework with contemporarymethods is done to demonstrate the increase in accuracy.

COVID19 Classification Using CT Images via Ensembles of Deep Learning Models

The recent COVID-19 pandemic caused by the novel coronavirus,severe acute respiratory syndrome coronavirus 2(SARS-CoV-2),has had a significant impact on human life and the economy around the world.A reverse transcription polymerase chain reaction(RT-PCR)test is used to screen for this disease,but its low sensitivity means that it is not sufficient for early detection and treatment.As RT-PCR is a time-consuming procedure,there is interest in the introduction of automated techniques for diagnosis.Deep learning has a key role to play in the field of medical imaging.The most important issue in this area is the choice of key features.Here,we propose a set of deep learning features based on a system for automated classification of computed tomography(CT)images to identify COVID-19.Initially,this method was used to prepare a database of three classes:Pneumonia,COVID19,and Healthy.The dataset consisted of 6000 CT images refined by a hybrid contrast stretching approach.In the next step,two advanced deep learning models(ResNet50 and DarkNet53)were fine-tuned and trained through transfer learning.The features were extracted from the second last feature layer of both models and further optimized using a hybrid optimization approach.For each deep model,the Rao-1 algorithm and the PSO algorithm were combined in the hybrid approach.Later,the selected features were merged using the new minimum parallel distance non-redundant(PMDNR)approach.The final fused vector was finally classified using the extreme machine classifier.The experimental process was carried out on a set of prepared data with an overall accuracy of 95.6%.Comparing the different classification algorithms at the different levels of the features demonstrated the reliability of the proposed framework.

Profiling Casualty Severity Levels of Road Accident Using Weighted Majority Voting

To determine the individual circumstances that account for a road traffic accident,it is crucial to consider the unplanned connections amongst various factors related to a crash that results in high casualty levels.Analysis of the road accident data concentrated mainly on categorizing accidents into different types using individually built classification methods which limit the prediction accuracy and fitness of the model.In this article,we proposed a multi-model hybrid framework of the weighted majority voting(WMV)scheme with parallel structure,which is designed by integrating individually implemented multinomial logistic regression(MLR)and multilayer perceptron(MLP)classifiers using three different accident datasets i.e.,IRTAD,NCDB,and FARS.The proposed WMV hybrid scheme overtook individual classifiers in terms of modern evaluation measures like ROC,RMSE,Kappa rate,classification accuracy,and performs better than state-of-theart approaches for the prediction of casualty severity level.Moreover,the proposed WMV hybrid scheme adds up to accident severity analysis through knowledge representation by revealing the role of different accident-related factors which expand the risk of casualty in a road crash.Critical aspects related to casualty severity recognized by the proposed WMV hybrid approach can surely support the traffic enforcement agencies to develop better road safety plans and ultimately save lives.

Skin Lesion Segmentation and Classification Using Conventional and Deep Learning Based Framework

Background:In medical image analysis,the diagnosis of skin lesions remains a challenging task.Skin lesion is a common type of skin cancer that exists worldwide.Dermoscopy is one of the latest technologies used for the diagnosis of skin cancer.Challenges:Many computerized methods have been introduced in the literature to classify skin cancers.However,challenges remain such as imbalanced datasets,low contrast lesions,and the extraction of irrelevant or redundant features.Proposed Work:In this study,a new technique is proposed based on the conventional and deep learning framework.The proposed framework consists of two major tasks:lesion segmentation and classification.In the lesion segmentation task,contrast is initially improved by the fusion of two filtering techniques and then performed a color transformation to color lesion area color discrimination.Subsequently,the best channel is selected and the lesion map is computed,which is further converted into a binary form using a thresholding function.In the lesion classification task,two pre-trained CNN models were modified and trained using transfer learning.Deep features were extracted from both models and fused using canonical correlation analysis.During the fusion process,a few redundant features were also added,lowering classification accuracy.A new technique called maximum entropy score-based selection(MESbS)is proposed as a solution to this issue.The features selected through this approach are fed into a cubic support vector machine(C-SVM)for the final classification.Results:The experimental process was conducted on two datasets:ISIC 2017 and HAM10000.The ISIC 2017 dataset was used for the lesion segmentation task,whereas the HAM10000 dataset was used for the classification task.The achieved accuracy for both datasets was 95.6% and 96.7%, respectively, which was higher thanthe existing techniques.

Citrus Diseases Recognition Using Deep Improved Genetic Algorithm

Agriculture is the backbone of each country,and almost 50%of the population is directly involved in farming.In Pakistan,several kinds of fruits are produced and exported the other countries.Citrus is an important fruit,and its production in Pakistan is higher than the other fruits.However,the diseases of citrus fruits such as canker,citrus scab,blight,and a few more impact the quality and quantity of this Fruit.The manual diagnosis of these diseases required an expert person who is always a time-consuming and costly procedure.In the agriculture sector,deep learning showing significant success in the last five years.This research work proposes an automated framework using deep learning and best feature selection for citrus diseases classification.In the proposed framework,the augmentation technique is applied initially by creating more training data from existing samples.They were then modifying the two pre-trained models named Resnet18 and Inception V3.The modified models are trained using an augmented dataset through transfer learning.Features are extracted for each model,which is further selected using Improved Genetic Algorithm(ImGA).The selected features of both models are fused using an array-based approach that is finally classified using supervised learning classifiers such as Support Vector Machine(SVM)and name a few more.The experimental process is conducted on three different datasets-Citrus Hybrid,Citrus Leaf,and Citrus Fruits.On these datasets,the best-achieved accuracy is 99.5%,94%,and 97.7%,respectively.The proposed framework is evaluated on each step and compared with some recent techniques,showing that the proposed method shows improved performance.

Human Gait Recognition Using Deep Learning and Improved Ant Colony Optimization

Human gait recognition(HGR)has received a lot of attention in the last decade as an alternative biometric technique.The main challenges in gait recognition are the change in in-person view angle and covariant factors.The major covariant factors are walking while carrying a bag and walking while wearing a coat.Deep learning is a new machine learning technique that is gaining popularity.Many techniques for HGR based on deep learning are presented in the literature.The requirement of an efficient framework is always required for correct and quick gait recognition.We proposed a fully automated deep learning and improved ant colony optimization(IACO)framework for HGR using video sequences in this work.The proposed framework consists of four primary steps.In the first step,the database is normalized in a video frame.In the second step,two pre-trained models named ResNet101 and InceptionV3 are selected andmodified according to the dataset’s nature.After that,we trained both modified models using transfer learning and extracted the features.The IACO algorithm is used to improve the extracted features.IACO is used to select the best features,which are then passed to the Cubic SVM for final classification.The cubic SVM employs a multiclass method.The experiment was carried out on three angles(0,18,and 180)of the CASIA B dataset,and the accuracy was 95.2,93.9,and 98.2 percent,respectively.A comparison with existing techniques is also performed,and the proposed method outperforms in terms of accuracy and computational time.

Video Analytics Framework for Human Action Recognition

Human action recognition(HAR)is an essential but challenging task for observing human movements.This problem encompasses the observations of variations in human movement and activity identification by machine learning algorithms.This article addresses the challenges in activity recognition by implementing and experimenting an intelligent segmentation,features reduction and selection framework.A novel approach has been introduced for the fusion of segmented frames and multi-level features of interests are extracted.An entropy-skewness based features reduction technique has been implemented and the reduced features are converted into a codebook by serial based fusion.A custom made genetic algorithm is implemented on the constructed features codebook in order to select the strong and wellknown features.The features are exploited by a multi-class SVM for action identification.Comprehensive experimental results are undertaken on four action datasets,namely,Weizmann,KTH,Muhavi,and WVU multi-view.We achieved the recognition rate of 96.80%,100%,100%,and 100%respectively.Analysis reveals that the proposed action recognition approach is efficient and well accurate as compare to existing approaches.

Smart Devices Based Multisensory Approach for Complex Human Activity Recognition

Sensors based Human Activity Recognition(HAR)have numerous applications in eHeath,sports,fitness assessments,ambient assisted living(AAL),human-computer interaction and many more.The human physical activity can be monitored by using wearable sensors or external devices.The usage of external devices has disadvantages in terms of cost,hardware installation,storage,computational time and lighting conditions dependencies.Therefore,most of the researchers used smart devices like smart phones,smart bands and watches which contain various sensors like accelerometer,gyroscope,GPS etc.,and adequate processing capabilities.For the task of recognition,human activities can be broadly categorized as basic and complex human activities.Recognition of complex activities have received very less attention of researchers due to difficulty of problem by using either smart phones or smart watches.Other reasons include lack of sensor-based labeled dataset having several complex human daily life activities.Some of the researchers have worked on the smart phone’s inertial sensors to perform human activity recognition,whereas a few of them used both pocket and wrist positions.In this research,we have proposed a novel framework which is capable to recognize both basic and complex human activities using builtin-sensors of smart phone and smart watch.We have considered 25 physical activities,including 20 complex ones,using smart device’s built-in sensors.To the best of our knowledge,the existing literature consider only up to 15 activities of daily life.