Search results

This work presents a system based on an ensemble of Convolutional Neural Networks (CNNs) and descriptors for bioimage classification that has been validated on different datasets of color images. The proposed system represents a very simple yet effective way of boosting the performance of trained CNNs by composing multiple CNNs into an ensemble and combining scores by sum rule. Several types of ensembles are considered, with different CNN topologies along with different learning parameter sets. The proposed system not only exhibits strong discriminative power but also generalizes well over multiple datasets thanks to the combination of multiple descriptors based on different feature types, both learned and handcrafted. Separate classifiers are trained for each descriptor, and the entire set of classifiers is combined by sum rule. Results show that the proposed system obtains state-of-the-art performance across four different bioimage and medical datasets. The MATLAB code of the descriptors will be available at https://github.com/LorisNanni.

Detecting cancer from the computed tomography (CT)images of lung nodules is very challenging for radiologists. Early detection of cancer helps to provide better treatment in advance and to enhance the recovery rate. Although a lot of research is being carried out to process clinical images, it still requires improvement to attain high reliability and accuracy. The main purpose of this paper is to achieve high accuracy in detecting and classifying the lung cancer and assisting the radiologists to detect cancer by using CT images. The CT images are collected from health-care centres and remote places through Internet of Things (IoT)-enabled platform and the image processing is carried out in the cloud servers.

IoT-based lung cancer detection is proposed to access the lung CT images from any remote place and to provide high accuracy in image processing. Here, the exact separation of lung nodule is performed by Otsu thresholding segmentation with the help of optimal characteristics and cuckoo search algorithm. The important features of the lung nodules are extracted by local binary pattern. From the extracted features, support vector machine (SVM) classifier is trained to recognize whether the lung nodule is malicious or non-malicious.

The proposed framework achieves 99.59% in accuracy, 99.31% in sensitivity and 71% in peak signal to noise ratio. The outcomes show that the proposed method has achieved high accuracy than other conventional methods in early detection of lung cancer.

The proposed algorithm is implemented and tested by using more than 500 images which are collected from public and private databases. The proposed research framework can be used to implement contextual diagnostic analysis.

The cancer nodules in CT images are precisely segmented by integrating the algorithms of cuckoo search and Otsu thresholding in order to classify malicious and non-malicious nodules.

Diagnosing pain in neonates is difficult but critical. Although approximately thirty manual pain instruments have been developed for neonatal pain diagnosis, most are complex, multifactorial, and geared toward research. The goals of this work are twofold: 1) to develop a new video dataset for automatic neonatal pain detection called iCOPEvid (infant Classification Of Pain Expressions videos), and 2) to present a classification system that sets a challenging comparison performance on this dataset. The iCOPEvid dataset contains 234 videos of 49 neonates experiencing a set of noxious stimuli, a period of rest, and an acute pain stimulus. From these videos 20 s segments are extracted and grouped into two classes: pain (49) and nopain (185), with the nopain video segments handpicked to produce a highly challenging dataset. An ensemble of twelve global and local descriptors with a Bag-of-Features approach is utilized to improve the performance of some new descriptors based on Gaussian of Local Descriptors (GOLD). The basic classifier used in the ensembles is the Support Vector Machine, and decisions are combined by sum rule. These results are compared with standard methods, some deep learning approaches, and 185 human assessments. Our best machine learning methods are shown to outperform the human judges.

Large amount of data are stored in image format. Image retrieval from bulk databases has become a hot research topic. An alternative method for efficient image retrieval is proposed based on a combination of texture and colour information. The main purpose of this paper is to propose a new content based image retrieval approach using combination of color and texture information in spatial and transform domains jointly.

Various methods are provided for image retrieval, which try to extract the image contents based on texture, colour and shape. The proposed image retrieval method extracts global and local texture and colour information in two spatial and frequency domains. In this way, image is filtered by Gaussian filter, then co-occurrence matrices are made in different directions and the statistical features are extracted. The purpose of this phase is to extract noise-resistant local textures. Then the quantised histogram is produced to extract global colour information in the spatial domain. Also, Gabor filter banks are used to extract local texture features in the frequency domain. After concatenating the extracted features and using the normalised Euclidean criterion, retrieval is performed.

The performance of the proposed method is evaluated based on the precision, recall and run time measures on the Simplicity database. It is compared with many efficient methods of this field. The comparison results showed that the proposed method provides higher precision than many existing methods.

The comparison results showed that the proposed method provides higher precision than many existing methods. Rotation invariant, scale invariant and low sensitivity to noise are some advantages of the proposed method. The run time of the proposed method is within the usual time frame of algorithms in this domain, which indicates that the proposed method can be used online.

Brain tumor segmentation and classification is the interesting area for differentiating the tumorous and the non-tumorous cells in the brain and classifies the tumorous cells for identifying its level. The methods developed so far lack the automatic classification, consuming considerable time for the classification. In this work, a novel brain tumor classification approach, namely, harmony cuckoo search-based deep belief network (HCS-DBN) has been proposed. Here, the images present in the database are segmented based on the newly developed hybrid active contour (HAC) segmentation model, which is the integration of the Bayesian fuzzy clustering (BFC) and the active contour model. The proposed HCS-DBN algorithm is trained with the features obtained from the segmented images. Finally, the classifier provides the information about the tumor class in each slice available in the database. Experimentation of the proposed HAC and the HCS-DBN algorithm is done using the MRI image available in the BRATS database, and results are observed. The simulation results prove that the proposed HAC and the HCS-DBN algorithm have an overall better performance with the values of 0.945, 0.9695 and 0.99348 for accuracy, sensitivity and specificity, respectively.

The proposed HAC segmentation approach integrates the properties of the AC model and BFC. Initially, the brain image with different modalities is subjected to segmentation with the BFC and AC models. Then, the Laplacian correction is applied to fuse the segmented outputs from each model. Finally, the proposed HAC segmentation provides the error-free segments of the brain tumor regions prevailing in the MRI image. The next step is to extract the useful features, based on scattering transform, wavelet transform and local Gabor binary pattern, from the segmented brain image. Finally, the extracted features from each segment are provided to the DBN for the training, and the HCS algorithm chooses the optimal weights for DBN training.

The experimentation of the proposed HAC with the HCS-DBN algorithm is analyzed with the standard BRATS database, and its performance is evaluated based on metrics such as accuracy, sensitivity and specificity. The simulation results of the proposed HAC with the HCS-DBN algorithm are compared against existing works such as k-NN, NN, multi-SVM and multi-SVNN. The results achieved by the proposed HAC with the HCS-DBN algorithm are eventually higher than the existing works with the values of 0.945, 0.9695 and 0.99348 for accuracy, sensitivity and specificity, respectively.

This work presents the brain tumor segmentation and the classification scheme by introducing the HAC-based segmentation model. The proposed HAC model combines the BFC and the active contour model through a fusion process, using the Laplacian correction probability for segmenting the slices in the database.

Happiness is an important mental emotion and yet becoming a major health concern nowadays. For this reason, better recognizing the objective understanding of how humans respond to event-related observations in their daily lives is especially important.

This paper uses non-intrusive technology (hyperspectral imaging [HSI]) for happiness recognition. Experimental setup is conducted for data collection in real-life environments where observers are showing spontaneous expressions of emotions (calm, happy, unhappy: angry) during the experimental process. Based on facial imaging captured from HSI, this work collects our emotional database defined as SWU Happiness DB and studies whether the physiological signal (i.e. tissue oxygen saturation [StO2], obtained by an optical absorption model) can be used to recognize observer happiness automatically. It proposes a novel method to capture local dynamic patterns (LDP) in facial regions, introducing local variations in facial StO₂ to fully use physiological characteristics with regard to hyperspectral patterns. Further, it applies a linear discriminant analysis-based support vector machine to recognize happiness patterns.

The results show that the best classification accuracy is 97.89 per cent, objectively demonstrating a feasible application of LDP features on happiness recognition.

This paper proposes a novel feature (i.e. LDP) to represent the local variations in facial StO₂ for modeling the active happiness. It provides a possible extension to the promising practical application.

The natural control, feedback, stimuli and protection of these subsequent principles founded this project. Via properly conducted experiments, a multilayer computer rehabilitation system was created that integrated natural interaction assisted by electroencephalogram (EEG), which enabled the movements in the virtual environment and real wheelchair. For blind wheelchair operator patients, this paper involved of expounding the proper methodology. For educating the value of life and independence of blind wheelchair users, outcomes have proven that virtual reality (VR) with EEG signals has that potential.

Individuals face numerous challenges with many disorders, particularly when multiple dysfunctions are diagnosed and especially for visually effected wheelchair users. This scenario, in reality, creates in a degree of incapacity on the part of the wheelchair user in terms of performing simple activities. Based on their specific medical needs, confined patients are treated in a modified method. Independent navigation is secured for individuals with vision and motor disabilities. There is a necessity for communication which justifies the use of VR in this navigation situation. For the effective integration of locomotion besides, it must be under natural guidance. EEG, which uses random brain impulses, has made significant progress in the field of health. The custom of an automated audio announcement system modified to have the help of VR and EEG for the training of locomotion and individualized interaction of wheelchair users with visual disability is demonstrated in this study through an experiment. Enabling the patients who were otherwise deemed incapacitated to participate in social activities, as the aim was to have efficient connections.

To protect their life straightaway and to report all these disputes, the military system should have high speed, more precise portable prototype device for nursing the soldier health, recognition of solider location and report about health sharing system to the concerned system. Field programmable gate array (FPGA)-based soldier’s health observing and position gratitude system is proposed in this paper. Reliant on heart rate which is centered on EEG signals, the soldier’s health is observed on systematic bases. By emerging Verilog hardware description language (HDL) programming language and executing on Artix-7 development FPGA board of part name XC7ACSG100t the whole work is approved in a Vivado Design Suite. Classification of different abnormalities and cloud storage of EEG along with the type of abnormalities, artifact elimination, abnormalities identification based on feature extraction, exist in the segment of suggested architecture. Irregularity circumstances are noticed through developed prototype system and alert the physically challenged (PHC) individual via an audio announcement. An actual method for eradicating motion artifacts from EEG signals that have anomalies in the PHC person’s brain has been established, and the established system is a portable device that can deliver differences in brain signal variation intensity. Primarily the EEG signals can be taken and the undesirable artifact can be detached, later structures can be mined by discrete wavelet transform these are the two stages through which artifact deletion can be completed. The anomalies in signal can be noticed and recognized by using machine learning algorithms known as multirate support vector machine classifiers when the features have been extracted using a combination of hidden Markov model (HMM) and Gaussian mixture model (GMM). Intended for capable declaration about action taken by a blind person, these result signals are protected in storage devices and conveyed to the controller. Pretending daily motion schedules allows the pretentious EEG signals to be caught. Aimed at the validation of planned system, the database can be used and continued with numerous recorded signals of EEG. The projected strategy executes better in terms of re-storing theta, delta, alpha and beta complexes of the original EEG with less alteration and a higher signal to noise ratio (SNR) value of the EEG signal, which illustrates in the quantitative analysis. The projected method used Verilog HDL and MATLAB software for both formation and authorization of results to yield improved results. Since from the achieved results, it is initiated that 32% enhancement in SNR, 14% in mean squared error (MSE) and 65% enhancement in recognition of anomalies, hence design is effectively certified and proved for standard EEG signals data sets on FPGA.

The proposed system can be used in military applications as it is high speed and excellent precise in terms of identification of abnormality, the developed system is portable and very precise. FPGA-based soldier’s health observing and position gratitude system is proposed in this paper. Reliant on heart rate which is centered on EEG signals the soldier health is observed in systematic bases. The proposed system is developed using Verilog HDL programming language and executing on Artix-7 development FPGA board of part name XC7ACSG100t and synthesised using in Vivado Design Suite software tool.

The defect detection problem of color-patterned fabric is still a huge challenge due to the lack of manual defect labeling samples. Recently, many fabric defect detection algorithms based on feature engineering and deep learning have been proposed, but these methods have overdetection or miss-detection problems because they cannot adapt to the complex patterns of color-patterned fabrics. The purpose of this paper is to propose a defect detection framework based on unsupervised adversarial learning for image reconstruction to solve the above problems.

The proposed framework consists of three parts: a generator, a discriminator and an image postprocessing module. The generator is able to extract the features of the image and then reconstruct the image. The discriminator can supervise the generator to repair defects in the samples to improve the quality of image reconstruction. The multidifference image postprocessing module is used to obtain the final detection results of color-patterned fabric defects.

The proposed framework is compared with state-of-the-art methods on the public dataset YDFID-1(Yarn-Dyed Fabric Image Dataset-version1). The proposed framework is also validated on several classes in the MvTec AD dataset. The experimental results of various patterns/classes on YDFID-1 and MvTecAD demonstrate the effectiveness and superiority of this method in fabric defect detection.

It provides an automatic defect detection solution that is convenient for engineering applications for the inspection process of the color-patterned fabric manufacturing industry. A public dataset is provided for academia.

This paper aims to propose a biologically inspired processing architecture to recognize and classify fabrics with respect to the weave pattern (fabric texture) and yarn color (fabric color).

By using the fabric weave patterns image identification system, this study analyzed the fabric image based on the Hierarchical-MAX (HMAX) model of computer vision, to extract feature values related to texture of fabric. Red Green Blue (RGB) color descriptor based on opponent color channels simulating the single opponent and double opponent neuronal function of the brain is incorporated in to the texture descriptor to extract yarn color feature values. Finally, support vector machine classifier is used to train and test the algorithm.

This two-stage processing architecture can be used to construct a system based on computer vision to recognize fabric texture and to increase the system reliability and accuracy. Using this method, the stability and fault tolerance (invariance) was improved.

Traditionally, fabric texture recognition is performed manually by visual inspection. Recent studies have proposed automatic fabric texture identification based on computer vision. In the identification process, the fabric weave patterns are recognized by the warp and weft floats. However, due to the optical environments and the appearance differences of fabric and yarn, the stability and fault tolerance (invariance) of the computer vision method are yet to be improved. By using our method, the stability and fault tolerance (invariance) was improved.

The traditional methods have difficulty to inspection various types of copper strips defects as inclusions, pits and delamination defects under uneven illumination. Therefore, this paper aims to propose an optimal real Gabor filter model for inspection; however, improper selection of Gabor parameters will cause the boundary between the defect and the background image to be not very clear. This will make the defect and the background cannot be completely separated.

The authors proposed an optimal Real Gabor filter model for inspection of copper surface defects under uneven illumination. This proposed method only requires a single filter by calculating the specific convolution energy of the Gabor filter with the image. The Real Gabor filter’s parameter is optimized by particle swarm optimization (PSO), which objective fitness function is maximization of the Gabor filter’s energy average divided by the energy standard deviation, the objective makes a distinction between the defect and normal area.

The authors have verified the effect with different iterations of parameter optimization using PSO, the effects with different control constant of energy and neighborhood window size of real Gabor filter, the experimental results on a number of metal surface have shown the proposed method achieved a well performance in defect recognition of metal surface.

The authors propose a defect detection method based on particle swarm optimization for single Gabor filter parameters optimization. This proposed method only requires a single filter and finds the best parameters of the Gabor filter. By calculating the specific convolution energy of the Gabor filter and the image, to obtain the best Gabor filter parameters and to highlight the defects, the particle swarm optimization algorithm’s fitness objective function is maximize the Gabor filter's average energy divided by the energy standard deviation.