Статьи журнала - International Journal of Image, Graphics and Signal Processing
Все статьи: 1092
Breast Cancer Classification from Ultrasound Images using VGG16 Model based Transfer Learning
Статья научная
Ultrasound based breast screening is gaining attention recently especially for dense breast. The technological advancement, cancer awareness, and cost-safety-availability benefits lead rapid rise of breast ultrasound market. The irregular shape, intensity variation, and additional blood vessels of malignant cancer are distinguishable in ultrasound images from the benign phase. However, classification of breast cancer using ultrasound images is a difficult process owing to speckle noise and complex textures of breast. In this paper, a breast cancer classification method is presented using VGG16 model based transfer learning approach. We have used median filter to despeckle the images. The layers for convolution process of the pretrained VGG16 model along with the maxpooling layers have been used as feature extractor and a proposed fully connected two layers deep neural network has been designed as classifier. Adam optimizer is used with learning rate of 0.001 and binary cross-entropy is chosen as the loss function for model optimization. Dropout of hidden layers is used to avoid overfitting. Breast Ultrasound images from two databases (total 897 images) have been combined to train, validate and test the performance and generalization strength of the classifier. Experimental results showed the training accuracy as 98.2% and testing accuracy as 91% for blind testing data with a reduced of computational complexity. Gradient class activation mapping (Grad-CAM) technique has been used to visualize and check the targeted regions localization effort at the final convolutional layer and found as noteworthy. The outcomes of this work might be useful for the clinical applications of breast cancer diagnosis.
Бесплатно
Breast lesion segmentation and area calculation for MR images
Статья научная
In this paper, our goal is to determine the boundaries of lesion and then calculate the area of existing lesion in breast magnetic resonance (MR) images to provide a useful information to the radiologists. For this purpose, at first stage region growing (RG) method and active contour model (Snake) is applied to the images to make the boundaries of lesion visible. RG method is one of the simplest approaches for image segmentation and provides accurate results with lower computation time due to its seed point initialization step. Snake method molds a closed contour to the boundary of a region in an image and is also popular in medical image segmentation studies. In the presented study, both of these methods are utilized to determine the lesion boundaries. After determining the boundaries of lesion accurately in the second stage of the study, bit-quad method is applied to the segmented images. Bit quad method is used to compute the area and perimeter of binary lesion images based on matching the logical state of regions of image to binary patterns. Finally, to evaluate the performance of the proposed study, computer simulations are performed. It is demonstrated via computer simulations that the lesion area and parameter values are very close to real values. By means of this study it is aimed to support radiologists during diagnosis and assessment of breast lesions.
Бесплатно
Статья научная
The growing quality and safety concern about fresh agricultural produce among consumers have led to the development of non-destructive quality assessment and testing techniques of fruits and vegetables. Humans judge the quality of fruits based on sensory attributes like taste, aroma etc. The shape, size, color, presence of defects which are external to fruits also influence the degree of consumer acceptability of produce. The traditional time consuming, manual fruit quality inspection is replaced by automated, fast, consistent, non-destructive techniques using computer vision in combination with learning algorithms. But the lack of benchmark datasets for agricultural produce has made an objective comparison of the proposed methods difficult. Hence, the proposed work aims to build a medium scale dataset for mango fruits of “Alphonso” cultivar with three classes: chilling injury, defective and non-defective. The reliability of the proposed dataset consisting of 2279 color images of mango fruits with 736 samples in chilling injury class, 632 samples in defective class and 911 samples in non-defective class, was established using a novel approach of developing a predictive model based on discriminant function analysis (DFA) which assigns group membership to each sample of the dataset. Extensive benchmarking analysis is established on the validated dataset using statistical and deep learning algorithms like support vector machine (SVM) and convolutional neural network (CNN), respectively. SVM achieved significant disease classification accuracy of 95% and 91.52% accuracy was achieved by custom CNN. The results of the proposed work indicate that the proposed color image dataset of mango fruits can be used as a benchmark dataset by other researchers for objective comparison in quality evaluation of mango fruits.
Бесплатно
COVID-19 Automatic Detection from CT Images through Transfer Learning
Статья научная
Identification of COVID-19 may help the community and patient to prevent the disease containment and plan to attend disease in right time. Deep neural network models widely used to analyze the medical images of COVID-19 for automatic detection and give the decision support for radiologists to summarize the accurate remarks. This paper proposed deep transfer learning for chest CT scan images to detection and diagnosis of COVID-19. VGG19, InceptionRestNetV3, InceptionV3 and DenseNet201 neural network used for automatic detection of COVID-19 disease form CT scan images (SARS-CoV-2 CT scan Dataset). Four deep transfer learning models were developed, tested and compared. The main objective of this paper is to use pre-trained features and converge pre-trained features with targeted features to improve the classification accuracy. It is observed that DenseNet201 noted the best performance and the classification accuracy is 99.98% for 300 epochs. The findings of the experiments show that the deeper networks struggle to train adequately and give less consistency when there is limited data. The DenseNet201 model adopted for COVID-19 identification from lung CT scans has been intensively optimized with optimal hyper parameters and performs at noteworthy levels with precision 99.2%, recall 100%, specificity 99.2%, and F1 score 99.2%.
Бесплатно
Статья научная
This work introduces the novelty as an application of histogram-based bins approach with statistical moments for detecting and classifying malaria using blood smear images into parasitized and uninfected cell images and the rising disease of COVID-19/Normal lung images. Proposed algorithms greatly vary as compared to the previous work. This work aims to improve accuracy in detection and classification and reduce feature vector dimensionality. It focuses on detailed image contents extracted into 8 bins by considering the significance of the R, G, and B color component relationship in the formation of each pixel. The texture features are represented by the first four moments for each of the three colors separately. This leads to the generation of 12 features vectors, each of size 8 components for each image in the database. Feature dimensionality reduction is achieved by applying different feature selection techniques to obtain desired optimum feature space. The comprehensive feature analysis presented here identifies many useful findings in order to validate the contribution of each image content uniquely in detection and classification. The proposed approach experimented with two image datasets: the malaria dataset obtained from the National Library of Medicine (NLM) and the lung image dataset acquired from the Radiography Database from Kaggle. The performance of work presented here is evaluated and compared with previous work with the same set of parameters, namely precision, recall, F1 score, and the AUC. We have achieved and improved the performances compared to previous work and also achieved better results even for the COVID-19 dataset.
Бесплатно
CRDL-PNet: An Efficient DeepLab-based Model for Segmenting Polyp Colonoscopy Images
Статья научная
Colorectal cancers are the third-largest kind of cancer in the world. However, detecting and removing precursor polyps with adenomatous cells using optical colonoscopy images helps to prevent this type of cancer. Moreover, hyperplastic polyps are benign cancers; adenomatous polyps are more likely to grow into cancerous tumors. Therefore, the detection and segmentation of polyps provide further histological evaluation. However, the main challenge is the extensive range of infected polyp features inside the colon and the lack of contrast between normal and infected areas. To solve these issues, the proposed novel Customized ResNet50 with DeepLabV3Plus Network (CRDL-PNet) model provided a scheme for segmenting polyps from colonoscopy images. The customized ResNet50 extracted features from polyp colonoscopy images. Furthermore, Atrous Spatial Pyramid Pooling (ASPP) is used to handle scale variation during training and improve feature selection maps in an upsampling layer. Additionally, the Gateaux Derivatives (GD) approach is used to segment boundary boxes of polyp regions. The proposed method has been evaluated on four datasets, namely the Kvasir-SEG, ETIS-PolypLaribDB, CVC-ClinicDB, and CVC-ColonDB datasets, for segmenting and detecting polyps. The simulation results have been examined by evaluation metrics, such as accuracy, Intersection-Over-Union (IOU), mean IOU, precision, recall, F1-score, dice, Jaccard, and Mean Process Time per Frame (MPTF) for proper validation. The proposed scheme outperforms the existing State-Of-The-Arts (SOTA) model on the same polyp datasets.
Бесплатно
Capsaicin-induced Thermal Enhancement on Target Tissues in Hyperthermia
Статья научная
Local thermal enhancement in target tissue is of great interest in tumor hyperthermia. In this study, we proposed a brand-new thermal enhancement protocol for tumor hyperthermia using heat generated from thermoge-nesis evoked by capsaicin, which can safely deliver a totally localized heating to target tissue. A healthy male volunteer was recruited, whose partial areas of the dorsum of hand and posterior aspect of forearm were smeared with 1% (w/w) capsaicin solution, to determine the increase of ther-mogenesis in human body. In addition, animal experiments on healthy Kunming (KM) mice (20-22g) were performed to test the feasibility and efficacy of capsaicin-induced thermal enhancement. These KM mice were first locally smeared with, subcutaneous or intraperitoneal injected of the same capsaicin solution, and then heated by near infrared laser. Preliminary experiments on the volunteer showed an effec-tive temperature increase in the skin area. Animal experi-ments indicated that distinct thermal enhancement in heat-ing effect, and that the thermal enhancement induced by intraperitoneal injection of capsaicin is more obvious than the other two ways. Thus capsaicin can be used as a poten-tial therapeutic adjuvant to locally enhance heating effects in target tissue during tumor hyperthermia.
Бесплатно
Central Moment and Multinomial Based Sub Image Clipped Histogram Equalization for Image Enhancement
Статья научная
The visual appearance of a digital image can be improved through image enhancement algorithm by reducing the noise in an image, improving the color, brightness and contrast of an image for more analysis. This paper introduces an image enhancement algorithm. The image histogram is processed through multinomial curvature fitting function to reduces the number of pixels for each intensity value through minimizing the sum of squared residuals. Then resampling is done to smooth out the computed data. After then histogram clipping threshold is computed by central moment processed on the resampled data value to restrict the over enhancement rate. Histogram is equally divided into two sub histograms. The sub histograms are equalized by transfer function to merged the sub images into one output image. The output image is further improved by reducing the environmental haze effect by applying Matlab imreducehaze method, which gives the final output image. Matlab simulation results demonstrate that the proposed method outperforms than other compared methods in terms of both quantitative and qualitative performance evaluation applied on colorfulness based PCQI (C-PCQI), and blind image quality measure of enhanced images (BIQME) image quality metrics.
Бесплатно
Change Energy Image for Gait Recognition: An Approach Based on Symbolic Representation
Статья научная
Gait can be identified by observing static and dynamic parts of human body. In this paper a variant of gait energy image called change energy images (CEI) are generated to capture detailed static and dynamic information of human gait. Radon transform is applied to CEI in four different directions (vertical, horizontal and two opposite cross sections) considering four different angles to compute discriminative feature values. The extracted features are represented in the form of interval –valued type symbolic data. The proposed method is capable of recognizing an individual when he/she have variations in their gait due to different clothes they wear, in different normal conditions and carrying a bag. A similarity measure suitable for the proposed gait representation is explored for the purpose of establishing similarity match for gait recognition. Experiments are conducted on CASIA database B and the results have shown better recognition performance compared to some of the existing methods.
Бесплатно
Channel Estimation of massive MIMO Using Code Shift Keying Pilot Symbols (CSK-PS)
Статья научная
The increasing demand for bandwidth by mobile users in wireless communication becomes a challenging issue to the research community. Several theories and models have been proposed to mitigate this issue. The most effective and commonly used approach to resolve the demand shortage of bandwidth is the massive Multi-Input and Multi-Output (MIMO) approach in which the number of transmitting and receiving antennas is placed at the base station (BS) to fulfill the issue of bandwidth. However, this technique suffers from various issues in estimating the channel due to interference, beamforming, and pilot contamination. In this paper, a novel channel estimation technique is being proposed using Code Shifting Keying symbols as pilot signals (CSK-PS) to minimize the pilot contamination. These signals are used as reference signals and the received signal is detected. The presented approach reduces the interference (pilot contamination) and improves the channel estimation in massive MIMO networks by using the modified expected propagation estimation method (MEPE). The presented approach is validated using mat-lab.
Бесплатно
Chaotic Behavior of Heart Rate Signals during Chi and Kundalini Meditation
Статья научная
Nonlinear dynamics has been introduced to the analysis of biological data and increasingly recognized to be functionally relevant. The aim of this study is to quantify and compare the contribution of nonlinear and chaotic dynamics of human heart rate variability during two forms of meditation: (i) Chinese Chi (or Qigong) meditation and (ii) Kundalini Yoga meditation. For this purpose, Poincare plots, Lyapunov exponents and Hurst exponents of heart rate variability signals were analyzed. In this study, we examined the different behavior of heart rate signals during two specific meditation techniques. The results show that heart rate signals became more periodic and their chaotic behavior was decreased in both techniques of meditation. Therefore, nonlinear chaotic indices may serve as a quantitative measure for psychophysiological states such as meditation. In addition, different forms of meditation appear to differentially alter specific components of heart rate signals.
Бесплатно
Chaotic Pixel Value Differencing
Статья научная
Pixel Value Differencing (PVD) is a spatial steganography technique that is area sensitive and considers complete visual invisibility while data hiding. While Least Significant Bit Approach (LSB) still remains the most popular technique and is simplest in approach its simplicity makes it vulnerable against steganalysis. Our proposed technique is an enhancement over traditional Pixel Value Differencing. We have added a layer of security using chaotic encryption approaches. Also some PVD based hybrid techniques are compared and analyzed to draw conclusions on the basis of various statistical measures.
Бесплатно
Chilli Dryness and Ripening Stages Assessment Using Machine Vision
Статья научная
The quality of chilli is prime concern for farmers, traders and chilli processing industries. The effective determination of chilli dryness and ripening stages are important factors in determining its quality and chilli shelf life with respect to manual estimation of ripening/dryness that are complex and time consuming. Chilli dryness and ripeness prediction at post-harvest stage by non-destructive machine vision technologies have potential of fair valuation for chilli produce for the chilli stalk holders. Chilli pericarp color values calculated from RGB, HSV and CIE-L*a*b* color space, texture properties using edge-wrinkles parameters are described by histogram of oriented gradients (HOG). LDA(linear discriminant analysis), RF(random-forest) and SVM(support vector machine) classifiers are analysed for performance accuracy for chilli dryness identification and chilli ripening stages using the machine vision. The chilli dryness identification accuracies of 83%, 85.4% and 83.5% are achieved using chilli color and HOG features with LDA, Random Forest and SVM classifiers respectively. Chilli ripening stage identification with combined chilli feature set of {color, HOG, SURF and LBP} using Support Vector Machine (SVM) average classifier accuracy is 90.56% across four chilli ripening stages. This work is simple with rapid, intelligent and high accuracy of chilli dryness and ripening identification by using machine vision approach has prospect in real-time chilli quality monitoring and grading. The results yielded were promising quality measurements compared previous studies.
Бесплатно
Classification and Recognition of Printed Hindi Characters Using Artificial Neural Networks
Статья научная
Character Recognition is one of the important tasks in Pattern Recognition. The complexity of the character recognition problem depends on the character set to be recognized. Neural Network is one of the most widely used and popular techniques for character recognition problem. This paper discusses the classification and recognition of printed Hindi Vowels and Consonants using Artificial Neural Networks. The vowels and consonants in Hindi characters can be divided in to sub groups based on certain significant characteristics. For each group, a separate network is designed and trained to recognize the characters which belong to that group. When a test character is given, appropriate neural network is invoked to recognize the character in that group, based on the features in that character. The accuracy of the network is analyzed by giving various test patterns to the system.
Бесплатно
Classification of Alzheimer Disease using Gabor Texture Feature of Hippocampus Region
Статья научная
The aim of this research is to propose a methodology to classify the subjects into Alzheimer disease and normal control on the basis of visual features from hippocampus region. All three dimensional MRI images were spatially normalized to the MNI/ICBM atlas space. Then, hippocampus region was extracted from brain structural MRI images, followed by application of two dimensional Gabor filter in three scales and eight orientations for texture computation. Texture features were represented on slice by slice basis by mean and standard deviation of magnitude of Gabor response. Classification between Alzheimer disease and normal control was performed with linear support vector machine. This study analyzes the performance of Gabor texture feature along each projection (axial, coronal and sagittal) separately as well as combination of all projections. The experimental results from both single projection (axial) as well as combination of all projections (axial, coronal and sagittal), demonstrated better classification performance over other existing method. Hence, this methodology could be used as diagnostic measure for the detection of Alzheimer disease.
Бесплатно
Classification of Brain Activity in Emotional States Using HOS Analysis
Статья научная
This paper proposes an emotion recognition system using EEG signals and higher order spectra. A visual induction based acquisition protocol is designed for recording the EEG signals in five channels (FP1, FP2, T3, T4 and Pz) under two emotional states of participants, calm-neutral and negatively exited. After pre-processing the signals, higher order spectra are employed to extract the features for classifying human emotions. We used Genetic Algorithm (GA) and Support vector machine (SVM) for optimum features selection for the classifier. In this research, we achieved an average accuracy of 82.32% for the two emotional states using Linear Discriminant Analysis (LDA) classifier. We concluded that, HOS analysis could be an accurate tool in the assessment of human emotional states. We achieved to same results compared to our previous studies.
Бесплатно
Статья научная
Epilepsy is a chronic brain disorder which affects normal neuronal activity of the brain. It results in sudden repeated episodes of higher electrical activity due to sensory disturbance. Electroencephalogram (EEG) plays an important role in the diagnosis of epilepsy. Currently, manual observation of EEG is done by experienced neurologist to diagnose epilepsy and related disorders. However, automated system is a promising method for seizure detection and diagnosis. The EEG signals recorded from the patient’s scalp are preprocessed, and classified as seizure and non-seizure based on the extracted signal features. The procedure significantly eliminates the error involved in manual observation. Due to non-linear nature of EEG, joint time-frequency methods are used to analyse the EEG signals. This paper proposes a EEG feature extraction technique using Dual Tree Complex Wavelet Transform (DTℂWT) to overcome the problem of shift variance in DWT. The estimation of improved multi-scale Permutation Entropy (IMPmEn) is done for the level-3 subband of DTℂWT. The performance of the Least Squares Support Vector Machine (LS-SVM) classifier is tested using these features and highest classification accuracy of 99.87 % is obtained on the real time EEG database.
Бесплатно
Classification of FNIRS Using Wigner-ville Distribution and CNN
Статья научная
Consumers undergo an intellectual burden when working with technological programs. Mostly in situations of several activities. For instance, while communicating when driving with the navigation device. It is not necessary to divert users from their primary duties in such circumstances. In memory cycles and related workload, the pre-frontal cortex (PFC) has a significant role to play. In this study, we have used data from 10 participants to evaluate the task behaviors in PFC with usable near-infrared spectroscopy (fNIRS), which is a non-invasive imaging modality. In classification, CNN research has been state of the art. This has undermined the need to extract features manually. In order to assess the mental workload, we implemented a time-frequency approach with CNN approach. Rather than traditional CNN network we used ResNet50 pretrained network here. Application of Wigner-Ville Distribution in Functional Imaging is introduced here. The proposed CNN approach achieves a considerable average improvement relative to conventional methods. The results across differences in time window length are benchmarked. Satisfactory result obtained with twenty five second window for which the CNN yields 98% correct classification where traditional CNN achieved 89% accuracy.
Бесплатно
Classification of High Blood Pressure Persons Vs Normal Blood Pressure Persons Using Voice Analysis
Статья научная
The human voice is remarkable, complex and delicate. All parts of the body play some role in voice production and may be responsible for voice dysfunction. The larynx contains muscles that are surrounded by blood vessels connected to circulatory system. The pressure of blood in these vessels should be related with dynamic variation of vocal cord parameters. These parameters are directly related with acoustic properties of speech. Acoustic voice analysis can be used to characterize the pathological voices. This paper presents the classification of high blood pressure and normal with the aid of voice signal recorded from the patients. Various features have been extracted from the voice signal of healthy persons and persons suffering from high blood pressure. Simulation results show differences in the parameter values of healthy and pathological persons. Then an optimum feature vector is prepared and kmean classification algorithm was implemented for data classification. The 79% classification efficiency was obtained.
Бесплатно
Classification of Mammogram Abnormalities Using Pseudo Zernike Moments and SVM
Статья научная
The most common malignancy observed among Indian women is the breast cancer. However, the cancer is detectable earlier by means of mammograms. Computer Aided Diagnostic (CAD) techniques are the boon to medical industry and these techniques intend to support the physicians in diagnosis. In this paper, a novel CAD system for the detection and classification of the abnormalities in the mammogram is presented. The proposed work is organized into four important phases and they are pre-processing, segmentation, feature extraction and classification. The pre-processing phase intends to remove unwanted noise and make the mammograms suitable for the next process. The segmentation phase aims to extract the areas of interest to proceed with further process. Feature extraction is the most important phase, which is meant for extracting the texture features from the area of interest. This work employs pseudo zernike moments for extracting features, owing to the noise resistance power and description ability. Finally, Support Vector Machine (SVM) is employed as the classifier, so as to distinguish between the malignant and normal mammograms. The performance of the proposed work is evaluated by several experimentations and the results are satisfactory in terms of accuracy, specificity and sensitivity.
Бесплатно
