Search

With the enormous increase in the use of the Internet, secure transfer of data across networks has become a challenging task. Attackers are in continuous search of getting information from network traffic, and this is the main reason that efficient intrusion detection techniques are required to identify different kinds of network attacks. In past, various supervised and semi-supervised methods have been developed for intrusion detection. Most of these methods require a large amount of data to develop an efficient intrusion detection system. In the proposed deep neural network and adaptive self-taught-based transfer learning technique, we have exploited the concept of self-taught learning to train deep neural networks for reliable network intrusion detection. In the proposed method, a pre-trained network on regression-related task is used to extract features from NSL-KDD dataset. Original features along with extracted features from the pre-trained network are then provided as an input to the sparse auto-encoder. Self-taught learning-based extracted features, when concatenated with the original features of NSL-KDD dataset, enhance the performance of the sparse auto-encoder. Performance of self-taught learning-based approach is compared against several techniques using ten independent runs in terms of accuracy, false alarm and detection rate, area under the ROC, and PR curve. It is experimentally observed that the auto-encoder trained on the combined original and extracted features is stable and offers good generalization in comparison with the sparse auto-encoder trained on original features alone.

The mitotic activity index is a key prognostic measure in tumour grading. Microscopy based detection of mitotic nuclei is a significant overhead and necessitates automation. This work proposes deep CNN based multi-phase mitosis detection framework “MP-MitDet” for mitotic nuclei identification in breast cancer histopathological images. The workflow constitutes: (1) label-refiner, (2) tissue-level mitotic region selection, (3) blob analysis, and (4) cell-level refinement. We developed an automatic label-refiner to represent weak labels with semi-sematic information for training of deep CNNs. A deep instance-based detection and segmentation model is used to explore probable mitotic regions on tissue patches. More probable regions are screened based on blob area and then analysed at cell-level by developing a custom CNN classifier “MitosRes-CNN” to filter false mitoses. The performance of the proposed “MitosRes-CNN” is compared with the state-of-the-art CNNs that are adapted to cell-level discrimination through cross-domain transfer learning and by adding task-specific layers. The performance of the proposed framework shows good discrimination ability in terms of F-score (0.75), recall (0.76), precision (0.71) and area under the precision-recall curve (0.78) on challenging TUPAC16 dataset. Promising results suggest good generalization of the proposed framework that can learn characteristic features from heterogenous mitotic nuclei.

In digital pathology, preliminary step for the development of Computer Aided Diagnostics involves defining and labelling an Object of Interest. Contrary to traditional tiresome method of observation and marking of objects under microscope, digital pathology enables the use of annotation tools to capture localization information and morphology for targeted object(s). However, it is quite difficult to capture this information due to high inter and intra class variance. In this paper, we have addressed these two different problems by developing an easy to use and efficient tool "Lymphocyte Annotator" that allows a user to draw bounding box(es) on IHC stained images, and as a result, automatically generates multiple variants of segmentation masks for the targeted object. Lymphocyte Annotator provides multiple options for bounding box selection, with variation in bounding box size and type. The developed tool targets labelling of lymphocytes in histopathological images. This annotator generates five different types of masks by using a) Chan Vese Segmentation Model, b) Morphological Active Contours without Edges Model, c) Morphological Geodesic Active Contours Model, d) Adaptive Thresholding and e) Circular Masks, thus providing user with an opportunity to use a mixture of different types of generated masks. Capturing of lymphocytes' morphology using different models incorporates morphological variance and enables training of robust deep Convolutional Neural Network. The annotations created from this tool are by default exported in PASCAL VOC and JSON format. Lymphocyte annotator is an open-source tool, and can be accessed from source, available at https://github.com/m-mohsin-zafar/lysto_labeltool.

Breast Cancer grading is a challenging task as regards image analysis, which is normally based on mitosis count rate. The mitotic count provides an estimate of aggressiveness of the tumor. The detection of mitosis is a challenging task because in a frame of slides at X40 magnification, there are hundreds of nuclei containing few mitotic nuclei. However, manual counting of mitosis by pathologists is a difficult and time intensive job, moreover conventional method rely mainly on the shape, color, and/or texture features as well as pathologist experience. The objective of this study is to accept the atypaia-2014 mitosis detection challenge, automate the process of mitosis detection and a proposal of a hybrid feature space that provides better discrimination of mitotic and non-mitotic nuclei by combining color features with morphological and texture features. To exploit color channels, they were first selected, and then normalized and cumulative histograms were computed in wavelet domain. A detailed analysis presented on these features in different color channels of respective color spaces using Random Forest (RF) and Support Vector Machine (SVM) classifiers. The proposed hybrid feature space when used with SVM classifier achieved a detection rate of 78.88% and F-measure of 72.07%. Our results, especially high detection rate, indicate that proposed hybrid feature space model contains discriminant information for mitotic nuclei, being therefore a very capable are for exploration to improve the quality of the diagnostic assistance in histopathology.

This paper presents a learning mechanism based on hybridization of static and dynamic learning. Realizing the detection performances offered by the state-of-the-art deep learning techniques and the competitive performances offered by the conventional static learning techniques, we propose the idea of exploitation of the concatenated (parallel) hybridization of the static and dynamic learning-based feature spaces. This is contrary to the cascaded (series) hybridization topology in which the initial feature space (provided by the conventional, static, and handcrafted feature extraction technique) is explored using deep, dynamic, and automated learning technique. Consequently, the characteristics already suppressed by the conventional representation cannot be explored by the dynamic learning technique. Instead, the proposed technique combines the conventional static and deep dynamic representation in concatenated (parallel) topology to generate an information-rich hybrid feature space. Thus, this hybrid feature space may aggregate the good characteristics of both conventional and deep representations, which are then explored using an appropriate classification technique. We also hypothesize that ensemble classification may better exploit this parallel hybrid perspective of the feature spaces. For this purpose, pyramid histogram of oriented gradients-based static learning has been incorporated in conjunction with convolution neural network-based deep learning to produce concatenated hybrid feature space. This hybrid space is then explored with various state-of-the-art ensemble classification techniques. We have considered the publicly available INRIA person and Caltech pedestrian standard image datasets to assess the performance of the proposed hybrid learning system. Furthermore, McNemar’s test has been used to statistically validate the outperformance of the proposed technique over various contemporary techniques. The validated experimental results show that the employment of the proposed hybrid representation results in effective detection performance (an AUC of 0.9996 for INRIA person and 0.9985 for Caltech pedestrian datasets) as compared to the individual static and dynamic representations.

Development of a diagnostic model using deep learning techniques is one of the eminent areas of research in computer aided diagnostic (CAD) systems. CAD systems based on deep learning have been developed to automate the process of mitosis classification. However, mitosis classification usually suffers from class imbalance problem, moreover, high similarity between mitotic and non-mitotic nuclei as well as discrepancy in appearance and shape make classification of mitotic nuclei a challenging area of research. Quality of CAD systems that use medical images for diagnosis, highly depends on how image features are manipulated in order to get better results. This work suggests a new technique of Channel Boosted Convolutional Neural Network (CB-CNN) to classify breast cancer mitotic nuclei. In this technique, feature representation of a CNN is boosted by adding auxiliary feature channels along with original feature space (Red, Green and Blue channels) to increase the generalization of model for heterogeneous and sparse data. In the proposed method, initially (80x80) patches of mitotic and non-mitotic nuclei were extracted from histopathological images by performing histogram based binary thresholding. For the development of a CB-CNN, the potential of auxiliary feature learners was exploited to learn high-level feature representation. In this regard, an additional set of texture and gradient based feature channels were concatenated with original RGB features space of data. This boosted representation was assigned to a custom made deep CNN model. Learning capacity of the proposed CB-CNN is evaluated on TUPAC'16 challenge dataset. Channel Boosting based CNNs (0.53 for CB-CNN, 0.71 for CB-VGG and 0.54 for CB-ResNet) show improved performance in terms of F-score as compared to SVM (F-score: 0.42), and baseline CNN (0.47 for CNN, 0.55 for VGG and 0.44 for ResNet) classifier with and without transfer learning. The improved performance of CB-CNN suggests that the boosting of channel representation improves the generalization of a CNN by making feature space more versatile and flexible, thus making it robust towards heterogeneous data.

The new emerging COVID-19, declared a pandemic disease, has affected millions of human lives and caused a massive burden on healthcare centers. Therefore, a quick, accurate, and low-cost computer-based tool is required to timely detect and treat COVID-19 patients. In this work, two new deep learning frameworks: Deep Hybrid Learning (DHL) and Deep Boosted Hybrid Learning (DBHL), is proposed for effective COVID-19 detection in X-ray dataset. In the proposed DHL framework, the representation learning ability of the two developed COVID-RENet-1 & 2 models is exploited individually through a machine learning (ML) classifier. In COVID-RENet models, Region and Edge-based operations are carefully applied to learn region homogeneity and extract boundaries features. While in the case of the proposed DBHL framework, COVID-RENet-1 & 2 are fine-tuned using transfer learning on the chest X-rays. Furthermore, deep feature spaces are generated from the penultimate layers of the two models and then concatenated to get a single enriched boosted feature space. A conventional ML classifier exploits the enriched feature space to achieve better COVID-19 detection performance. The proposed COVID-19 detection frameworks are evaluated on radiologist's authenticated chest X-ray data, and their performance is compared with the well-established CNNs. It is observed through experiments that the proposed DBHL framework, which merges the two-deep CNN feature spaces, yields good performance (accuracy: 98.53%, sensitivity: 0.99, F-score: 0.98, and precision: 0.98). Furthermore, a web-based interface is developed, which takes only 5–10s to detect COVID-19 in each unseen chest X-ray image. This web-predictor is expected to help early diagnosis, save precious lives, and thus positively impact society., Graphical abstract Image 1

Background The recent emergence of a highly infectious and contagious respiratory viral disease known as COVID-19 has vastly impacted human lives and overloaded the health care system. Therefore, it is crucial to develop a fast and accurate diagnostic system for the timely identification of COVID-19 infected patients and thus to help control its spread. Methods This work proposes a new deep CNN based technique for COVID-19 classification in X-ray images. In this regard, two novel custom CNN architectures, namely COVID-RENet-1 and COVID-RENet-2, are developed for COVID-19 specific pneumonia analysis. The proposed technique systematically employs Region and Edge-based operations along with convolution operations. The advantage of the proposed idea is validated by performing series of experimentation and comparing results with two baseline CNNs that exploited either a single type of pooling operation or strided convolution down the architecture. Additionally, the discrimination capacity of the proposed technique is assessed by benchmarking it against the state-of-the-art CNNs on radiologist's authenticated chest X-ray dataset. Implementation is available at https://github.com/PRLAB21/Coronavirus-Disease-Analysis-using-Chest-X-Ray-Images . Results The proposed classification technique shows good generalization as compared to existing CNNs by achieving promising MCC (0.96), F-score (0.98) and Accuracy (98%). This suggests that the idea of synergistically using Region and Edge-based operations aid in better exploiting the region homogeneity, textural variations, and region boundary-related information in an image, which helps to capture the pneumonia specific pattern. Conclusions The encouraging results of the proposed classification technique on the test set with high sensitivity (0.98) and precision (0.98) suggest the effectiveness of the proposed technique. Thus, it suggests the potential use of the proposed technique in other X-ray imagery-based infectious disease analysis.

Mitotic nuclei estimation in breast tumour samples has a prognostic significance in analysing tumour aggressiveness and grading system. The automated assessment of mitotic nuclei is challenging because of their high similarity with non-mitotic nuclei and heteromorphic appearance. In this work, we have proposed a new Deep Convolutional Neural Network (CNN) based Heterogeneous Ensemble technique “DHE-Mit-Classifier” for analysis of mitotic nuclei in breast histopathology images. The proposed technique in the first step detects candidate mitotic patches from the histopathological biopsy regions, whereas, in the second step, these patches are classified into mitotic and non-mitotic nuclei using the proposed DHE-Mit-Classifier. For the development of a heterogeneous ensemble, five different deep CNNs are designed and used as base-classifiers. These deep CNNs have varying architectural designs to capture the structural, textural, and morphological properties of the mitotic nuclei. The developed base-classifiers exploit different ideas, including (i) region homogeneity and feature invariance, (ii) asymmetric split-transform-merge, (iii) dilated convolution based multi-scale transformation, (iv) spatial and channel attention, and (v) residual learning. Multi-layer-perceptron is used as a meta-classifier to develop a robust and accurate classifier for providing the final decision. The performance of the proposed ensemble “DHE-Mit-Classifier” is evaluated against state-of-the-art CNNs. The performance evaluation on the test set suggests the superiority of the proposed ensemble with an F-score (0.77), recall (0.71), precision (0.83), and area under the precision-recall curve (0.80). The good generalisation of the proposed ensemble with a considerably high F-score and precision suggests its potential use in the development of an assistance tool for pathologists.

Breast cancer is one of the deadly diseases among women. However, the chances of death are highly reduced if it gets diagnosed and treated at its early stage. Mammography is one of the reliable methods used by the radiologist to detect breast cancer at its initial stage. Therefore, an automatic and secure breast cancer detection system that accurately detects abnormalities not only increases the radiologist's diagnostic confidence but also provides more objective evidence. In this work, an automatic Diverse Features based Breast Cancer Detection (DFeBCD) system is proposed to classify a mammogram as normal or abnormal. Four sets of distinct feature types are used. Among them, features based on taxonomic indexes, statistical measures and local binary patterns are static. The proposed DFeBCD dynamically extracts the fourth set of features from mammogram images using a highway-network based deep convolution neural network (CNN). Two classifiers, Support Vector Machine (SVM) and Emotional Learning inspired Ensemble Classifier (ELiEC), are trained on these distinct features using a standard IRMA mammogram dataset. The reliability of the system performance is ensured by applying 5-folds cross-validation. Through experiments, we have observed that the performance of the DFeBCD system on dynamically generated features through highway network-based CNN is better than that of all the three individual sets of ad-hoc features. Furthermore, the hybridization of all four types of features improves the system's performance by nearly 2-3%. The performance of both the classifiers is comparable using the individual sets of ad-hoc features. However, the ELiEC classifier's performance is better than SVM using both hybrid and dynamic features.

Display Omitted An intelligent reversible watermarking technique for the protection of medical images.The concept of companding function is exploited for reducing embedding distortion.Integer Wavelet Transform is used as an embedding domain for achieving reversibility.Histogram processing is employed to avoid underflow/overflow.GP is used to evolve models that not only make an optimal tradeoff between imperceptibility and capacity of watermark. This work presents an intelligent technique based on reversible watermarking for protecting patient and medical related information. In the proposed technique IRW-Med, the concept of companding function is exploited for reducing embedding distortion, while Integer Wavelet Transform (IWT) is used as an embedding domain for achieving reversibility. Histogram processing is employed to avoid underflow/overflow. In addition, the learning capabilities of Genetic Programming (GP) are exploited for intelligent wavelet coefficient selection. In this context, GP is used to evolve models that not only make an optimal tradeoff between imperceptibility and capacity of the watermark, but also exploit the wavelet coefficient hidden dependencies and information related to the type of sub band. The novelty of the proposed IRW-Med technique lies in its ability to generate a model that can find optimal wavelet coefficients for embedding, and also acts as a companding factor for watermark embedding. The proposed IRW-Med is thus able to embed watermark with low distortion, take out the hidden information, and also recovers the original image. The proposed IRW-Med technique is effective with respect to capacity and imperceptibility and effectiveness is demonstrated through experimental comparisons with existing techniques using standard images as well as a publically available medical image dataset.

Prediction system "IEH-GT" for protein localization in fluorescence microscopy images.Utilizing information about the patterns of intensity-variations of proteins.New feature extraction based on individual features along various orientations.Discriminative learning through individual feature spaces of both GLCMs and textons.Employing the learning capabilities of ensemble using SVM and majority voting. Image based subcellular localization of proteins is a challenging pattern recognition task and is of prime importance in understanding different functions of proteins. Extracting information about the protein's subcellular location using discriminative feature extraction strategies helps in understanding the behavior of different proteins under different circumstances.In this work, an intelligent prediction system for protein subcellular localization using fluorescence microscopy images is developed. The proposed prediction system IEH-GT uses a new feature extraction strategy and ensemble classification. The new feature extraction mechanism exploits statistical and Texton based image descriptors, whereas ensemble classification is performed using the majority voting based ensemble of SVMs. The novelty of IEH-GT prediction system lies (a) in the individual exploitation of the feature spaces generated from both individual Gray Level Co-occurrence Matrices (GLCMs) and Texton images as well as (b) in the manner the extracted features are exploited using the learning capabilities of SVM, which is utilized as base classifier in majority voting based ensemble. In this work, the effect of the individual orientations of GLCMs and Texton images on the capability of protein-recognition is investigated.It is observed that GLCMs along different orientations characterize different information regarding intensity distribution and relative positions of pixels in a certain neighborhood. Similarly, the individually generated features from each Texton image help in reducing the overlapping information. The proposed IEH-GT prediction system is thus able to effectively exploit information about the patterns of intensity-variations of proteins. It is experimentally observed that due to the capability of the proposed IEH-GT system in differentiating diverse surface structures of proteins, the mis-predictions of similar proteins such as Golgpp and Golgia are substantially reduced.

Automating the scoring of Whole-Slide Images (WSIs) is a challenging task because the search space for selecting region of interest (ROI) is huge due to the very large sizes of WSIs. A Multifaceted Fused-CNN (MF-CNN) and a Hybrid-Descriptor are proposed to develop an integrated scoring system for Breast Cancer histopathology WSIs. Suitable color and textural features are identified to help mitotic count based selection of ROIs at lower resolution. To recognize complex patterns, the MF-CNN considers multiple facets of the input image. It counts mitoses, extracts handcrafted features from ROIs and utilizes global texture of the images to form a Hybrid-Descriptor for training a classifier assigning scores to WSIs. The proposed system is evaluated on a publicly available benchmark (TUPAC16) and produced the highest score of 0.582 in terms of Cohen’s Kappa. It surpassed human experts’ level accuracy of ROI selection and can therefore reduce the burden of manual ROI selection for WSIs.

Segmentation and detection of mitotic nuclei is a challenging task. To address this problem, a Transfer Learning based fast and accurate system is proposed. To give the classifier a balanced dataset, this work exploits the concept of Transfer Learning by first using a pre-trained convolutional neural network (CNN) for segmentation, and then another Hybrid-CNN (with Weights Transfer and custom layers) for classification of mitoses. First, mitotic nuclei are automatically annotated, based on the ground truth centroids. The segmentation module then segments mitotic nuclei and also produces some false positives. Finally, the detection module is trained on the patches from the segmentation module and performs the final detection. Fine-tuning based Transfer Learning reduced training time, provided good initial weights, and improved the detection rate with F-measure of 0.713 and 76% area under the precision-recall curve for the challenging task of mitosis detection.

This study presents the analysis of hepatitis B virus (HBV) infection in human blood serum using Raman spectroscopy combined with pattern recognition technique. In total, 119 confirmed samples of HBV infected sera, collected from Pakistan Atomic Energy Commission (PAEC) general hospital have been used for the current analysis. The differences between normal and HBV infected samples have been evaluated using support vector machine (SVM) algorithm. SVM model with two different kernels i.e. polynomial function and Gaussian radial basis function (RBF) have been investigated for the classification of normal blood sera from HBV infected sera based on Raman spectral features. Furthermore, the performance of the model with each kernel function has also been analyzed with two different implementations of optimization problem i.e. Quadratic programming and least square. 5-fold cross validation method has been used for the evaluation of the model. In the current study, best classification performance has been achieved for polynomial kernel of order-2. A diagnostic accuracy of about 98% with the precision of 97%, sensitivity of 100% and specificity of 95% has been achieved under these conditions.

In this paper, a hierarchal feature extraction and ensemble classification-based framework for object detection is proposed. The proposed object detection technique is motivated by the hierarchical learning mechanism in primate visual cortex, where each layer processes information differently. Initially, pyramid histogram of oriented gradients (PHOG) based descriptors are selected to generate shift and scale invariant descriptors of an image. PHOG-based feature descriptors are then processed in multi-layered hierarchy, following feed forward models in brain's visual cortex and exploited through ensemble classification techniques. The proposed cortex-inspired ensemble-based object detection (CI-EnsOD) system exploits hierarchical learning mechanism of visual cortex and it is computationally efficient compared to the existing cortex-inspired models. In addition, it reduces feature dimensionality and offers improved object detection performance. The performance of proposed technique is demonstrated using three publically available standard datasets. It is shown experimentally that the prototype selection in the proposed CI-EnsOD can be improved using k-means clustering. The obtained experimental results show that the proposed CI-EnsOD technique is more accurate and efficient than contemporary cortex-inspired object detection techniques. Finally, it is also observed that the proposed technique is capable of providing compact descriptors compared to principle component analysis and independent component analysis.

This work presents the evaluation of Raman spectroscopy using random forest (RF) for the analysis of dengue fever in the infected human sera. A total of 100 dengue suspected blood samples, collected from Holy Family Hospital, Rawalpindi, Pakistan, have been used in this study. Out of these samples, 45 were dengue-positive based on immunoglobulin M (IgM) capture enzyme-linked immunosorbent assay (ELISA) tests. For highlighting the spectral differences between normal and infected samples, an effective machine learning system is developed that automatically learns the pattern of the shift in spectrum for the dengue compared to normal cases and thus is able to predict the unknown class based on the known example. In this connection, dimensionality reduction has been performed with the principal component analysis (PCA), while RF is used for automatic classification of dengue samples. For the determination of diagnostic capabilities of Raman spectroscopy based on RF, sensitivity, specificity, and accuracy have been calculated in comparison to normally performed IgM capture ELISA. According to the experiment, accuracy of 91%, sensitivity of 91%, and specificity of 91% were achieved for the proposed RF-based model.

Anomaly detection in a network is one of the prime concerns for network security. In this work, a novel Channel Boosted and Residual learning based deep Convolutional Neural Network (CBR-CNN) architecture is proposed for the detection of network intrusions. The proposed methodology is based on inherent nature of the anomaly detection in which one class classification approach is used to detect network intrusion. This is accomplished by the modelling of normal network traffic distribution using Stacked Autoencoders (SAE). Using unsupervised training, SAE transforms the original feature space into a reconstructed feature space, which is further transformed via the proposed concept of channel boosting. Additionally, in order to increase the representational power of the neural network and the diversity in features representation, a multipath residual learning based CNN architecture is proposed to learn features at different levels of granularity. Performance of the proposed CBR-CNN technique is evaluated on NSL-KDD dataset. Our proposed method showed significant improvement over the existing techniques, achieving accuracy, AU-ROC, and AU-PR of 89.41%, 0.9473, and 0.9443 on Test + and 80.36%, 0.7348 and 0.9034 on Test−21 dataset, respectively.

A novel prediction error expansion-based reversible image watermarking technique is proposed. The proposed approach exploits the variance of neighbouring pixels in a way that the embedding of information bits in each pixel of an image depends on the scale of variation present in its eight neighbouring pixel values. A rhombus predictor is employed in the proposed scheme for its efficacy. In contrast to existing prediction-error expansion based reversible watermarking techniques, the embedding phase of the proposed approach depends only on the context of the pixel to be watermarked. Exploiting the neighbourhood information of each pixel in this way, avoids the need of a location map (LM) or histogram shifting (HS) to distinguish marked pixels from unmarked pixels. Consequently, the proposed approach does not require a large amount of auxiliary information in the form of a LM or HS of unmarked pixels, and thus introduces less distortion, yielding good embedding performance, especially at high capacity.

Amongst the various data hiding schemes, lossless data hiding has achieved substantial attention in the recent years. Lossless watermarking also known as reversible watermarking, not only ensures the extraction of the hidden information, called watermark, but also the complete recovery of the original image as well. Recently, several interesting reversible block-based watermarking approaches using histogram processing have been reported. These approaches offer substantial watermarking capacity for a fixed image quality measure. However, there is considerable margin of improvement as regards the imperceptibility versus capacity tradeoff is concerned. Some of the watermarking applications such as those related to integrity-control need high capacity. Therefore, we present a high capacity reversible watermarking approach based on histogram processing and block selection (RW-HPBS). The proposed RW-HPBS approach exploits the concept of down sampling for effectively increasing the capacity and is useful for integrity control and authentication related applications. Two down-sampled versions, the reference and the data hiding, are obtained to create space for embedding. Authentication property is achieved by taking a secondary advantage of the location map before being compressed and embedded as an auxiliary information. The effectiveness of the proposed method is shown through comparison with the existing techniques. Experimental results show that the proposed RW-HPBS approach provides better imperceptibility versus capacity tradeoff and can detect tampering and collage attack.

This paper presents a framework based on a single dual-purpose semi-fragile watermark to verify the integrity of digital image along with the recovery of distorted image. The watermark is correlated to the host image for detecting the collage attack and then embedded in their respective wavelet subbands. Unlike the conventional block-based approaches, this work has the ability to determine the unverified regions concisely. Huffman and BCH coding are utilized while generating the watermark. Integer DCT has been exploited as it can be highly compressed by Huffman coding as compared to the conventional DCT contents. The proposed technique exhibits the flexibility between imperceptibility, robustness, and capacity. In addition, integer wavelet transform has been used to reduce the computational complexity of the algorithm. Evaluation of experimental investigation shows the performance of dual-purpose semi-fragile watermark.

Mitochondrial protein of Plasmodium falciparum is an important target for anti-malarial drugs. Experimental approaches for detecting mitochondrial proteins are costly and time consuming. Therefore, MitProt-Pred is developed that utilizes Bi-profile Bayes, Pseudo Average Chemical Shift, Split Amino Acid Composition, and Pseudo Amino Acid Composition based features of the protein sequences. Hybrid feature space is also developed by combining different individual feature spaces. These feature spaces are learned and exploited through SVM based ensemble. MitProt-Pred achieved significantly improved prediction performance for two standard datasets. We also developed the score level ensemble, which outperforms the feature level ensemble.

In this paper, we present a hybrid, image restoration approach. The proposed approach combines the geostatistical interpolation of punctual kriging, artificial neural networks (ANNs), and fuzzy logic based approaches. Images degraded with Gaussian white noise are restored by first utilizing fuzzy logic for selecting pixels that needs kriging. Three fuzzy systems are employed. Both type-I and type-II fuzzy sets in addition with neuro fuzzy classifier (NFC) have been used for the detection of noisy pixels. To avoid edge pixels, a post processing technique is used to check the edge pixel connectivity up to lag 5. If the pixel under consideration is an edge pixel, it is excluded from the fuzzy map and thus not estimated. The concept of punctual kriging is then used to estimate the intensity of a noisy pixel. ANN is employed to minimize the cost function of the kriging based pixel intensity estimation procedure. ANN, in contrast to analytical methodologies, avoids both matrix inversion failure and negative weights problems. Image restoration performance based comparison has been made against adaptive Weiner filter and existing fuzzy kriging approaches. Experimental results using 450 images are used to validate the effectiveness of the proposed approach. Different image quality measures are used to compare the efficacy of the proposed NFC and fuzzy type-II approaches for detecting noisy pixels in conjunction with ANN and kriging based estimation.

Different types of breast cancer are affecting lives of women across the world. Common types include Ductal carcinoma in situ (DCIS), Invasive ductal carcinoma (IDC), Tubular carcinoma, Medullary carcinoma, and Invasive lobular carcinoma (ILC). While detecting cancer, one important factor is mitotic count - showing how rapidly the cells are dividing. But the class imbalance problem, due to the small number of mitotic nuclei in comparison to the overwhelming number of non-mitotic nuclei, affects the performance of classification models. This work presents a two-phase model to mitigate the class biasness issue while classifying mitotic and non-mitotic nuclei in breast cancer histopathology images through a deep convolutional neural network (CNN). First, nuclei are segmented out using blue ratio and global binary thresholding. In Phase-1 a CNN is then trained on the segmented out 80×80 pixel patches based on a standard dataset. Hard non-mitotic examples are identified and augmented; mitotic examples are oversampled by rotation and flipping; whereas non-mitotic examples are undersampled by blue ratio histogram based k-means clustering. Based on this information from Phase-1, the dataset is modified for Phase-2 in order to reduce the effects of class imbalance. The proposed CNN architecture and data balancing technique yielded an F-measure of 0.79, and outperformed all the methods relying on specific handcrafted features, as well as those using a combination of handcrafted and CNN-generated features.

Membrane proteins are fundamental elements of a cell that play essential roles nearly in all the cellular processes. Prediction of membrane protein types using biological experiments are often complicated and time consuming. Therefore it is highly desirable to develop a robust, reliable and high-throughput silico method to predict membrane protein types. In this study, the authors have used two feature extraction strategies known as dipeptide and pseudo amino acid (PseAA) compositions for classification of membrane proteins types. In addition, a composite model is also developed by concatenating dipeptide and PseAA composition based features. Further, two feature selection methods such as neighbourhood preserving embedding and locally linear embedding (LLE) are applied to reduce the dimensionality of the composite model. The performance of these feature extraction strategies is evaluated using four different classifiers: K-nearest neighbour, probabilistic neural network (PNN), support vector machine (SVM) and grey incidence degree. The highest success rates have been observed using the LLE-based reduced features. SVM has yielded the best accuracy of 88.2% in case of jackknife test. Although in case of independent dataset test, PNN has obtained the highest accuracy of 98.4%. Performance measures other than accuracy are also used such as ‘Mathew correlation coefficient’, sensitivity and precision. The authors simulated results show that the composite model has significantly discriminated the types of membrane protein and might be useful for future research and drug discovery.

This letter presents a novel approach of incorporating intelligence in the encoding and decoding structures of a watermarking system. The employment of computational intelligence makes the watermarking system resistant against a series of attacks, which may occur during the storage or communication of the watermarked work. Keeping in view the Human Visual System, Genetic Programming is used to generate functions which select optimum strength and location of transform domain coefficients for watermark embedding. Support Vector Machines and Artificial Neural Networks are employed at the decoding side to learn about the distortions due to attacks and counteract them. Especially, the proposed system is quite effective for robust watermarking applications of small size images such as those displayed on portable devices and online catalogs.

The use of intelligent and sophisticated approaches in the domain of computer vision and pattern recognition is consistently increasing. This paper introduces a novel machine learning based approach for Shape From Focus (SFF), where the in-focus pixels are selected from a sequence of images. In contrast to computing focus values directly in spatial or transform domain, the proposed method first nonlinearly transforms the input space into feature space and then computes the focus value by transforming the data into eigenspace. First, the nonlinear transformation is performed by using kernel function and then Principal Component Analysis (PCA) is applied. This idea is also supported by the fact that out-of-focus is analogous to blurring and is a nonlinear phenomenon. An initial depth map is computed by maximizing the focus measure. To further refine the 3D shape, bilateral filter is applied. The proposed method is experimented using synthetic and real image sequences. The results demonstrate the effectiveness and the robustness of the new method.

This work presents a diagnostic system for the hepatitis C infection using Raman spectroscopy and proximity based classification. The proposed method exploits transformed Raman spectra using the proximity based machine learning technique and is denoted as RS-PCA-Prox. First, Raman spectral data is baseline corrected by subtracting noise and low intensity background. After this, a feature transformation of Raman spectra is adopted, not only to reduce the feature’s dimensionality but also to learn different deviations in Raman shifts. The proposed RS-PCA-Prox shows significant diagnostic power in terms of accuracy, sensitivity, and specificity as 95%, 0.97 and 0.94 in PCA based transformed domain. The comparison of the RS-PCA-Prox with linear and ensemble based classifiers shows that proximity based classification performs better for the discrimination of HCV infected individuals and is able to differentiate the infected individuals from normal ones on the basis of molecular spectral information. Furthermore, it is observed that characteristic spectral changes are due to variation in the intensity of lectin, chitin, lipids, ammonia and viral protein as a consequence of the HCV infection.

This paper presents a novel approach of adaptive visual tuning of a watermark in Discrete Cosine Transform (DCT) domain. The proposed approach intelligently selects appropriate frequency bands as well as optimal strength of alteration. Genetic Programming (GP) is applied to structure the watermark by exploiting both the characteristics of human visual system and information pertaining to a cascade of conceivable attacks. The developed visual tuning expressions are dependent on frequency and luminance sensitivities, and contrast masking. To further enhance robustness, spread spectrum based watermarking and Bose-Chadhuri-Hocquenghem (BCH) coding is employed. The combination of spread spectrum sequence, BCH coding and GP based non-linear structuring makes it extremely difficult for an attacker to gain information about the secret knowledge of the watermarking system. Experimental results show the superiority of the proposed approach against the existing approaches. Especially, the margin of improvement in robustness will be of high importance in medical and context aware related applications of watermarking.

We present an innovative scheme of blindly extracting message bits when a watermarked image is distorted. In this scheme, we have exploited the capabilities of machine learning (ML) approaches for nonlinearly classifying the embedded bits. The proposed technique adaptively modifies the decoding strategy in view of the anticipated attack. The extraction of bits is considered as a binary classification problem. Conventionally, a hard decoder is used with the assumption that the underlying distribution of the discrete cosine transform coefficients do not change appreciably. However, in case of attacks related to real world applications of watermarking, such as JPEG compression in case of shared medical image warehouses, these coefficients are heavily altered. The sufficient statistics corresponding to the maximum likelihood based decoding process, which are considered as features in the proposed scheme, overlap at the receiving end, and a simple hard decoder fails to classify them properly. In contrast, our proposed ML decoding model has attained highest accuracy on the test data. Experimental results show that through its training phase, our proposed decoding scheme is able to cope with the alterations in features introduced by a new attack. Consequently, it achieves promising improvement in terms of bit correct ratio in comparison to the existing decoding scheme.

We present a novel approach to developing Machine Learning (ML) based decoding models for extracting a watermark in the presence of attacks. Statistical characterization of the components of various frequency bands is exploited to allow blind extraction of the watermark. Experimental results show that the proposed ML based decoding scheme can adapt to suit the watermark application by learning the alterations in the feature space incurred by the attack employed.

In this work, we predict G-protein coupled receptors (GPCRs) using hydrophobicity of amino acid sequences and Fast Fourier Transform for feature generation. We analyze whether the GPCRs classification strategy depends on the way the feature space may be exploited. Consequently, we show that the sequence pattern based information could easily be exploited in the frequency domain using proximity rather than increasing margin of separation between the classes. We thus develop a simple proximity based approach known as nearest neighbor (NN) for classifying the 17 GPCRs subfamilies. The NN classifier has outperformed the one against all implementation of support vector machine using both Jackknife and independent dataset. The results validate the importance of the understanding and efficient exploitation of the feature space. It also shows that simple classification strategies may outperform complex ones because of the efficient exploitation of the feature space.

Gene expression deviates from its normal composition in case a patient has cancer. This variation can be used as an effective tool to find cancer. In this study, we propose a novel gene expressions based colon classification scheme (GECC) that exploits the variations in gene expressions for classifying colon gene samples into normal and malignant classes. Novelty of GECC is in two complementary ways. First, to cater overwhelmingly larger size of gene based data sets, various feature extraction strategies, like, chi-square, F-Score, principal component analysis (PCA) and minimum redundancy and maximum relevancy (mRMR) have been employed, which select discriminative genes amongst a set of genes. Second, a majority voting based ensemble of support vector machine (SVM) has been proposed to classify the given gene based samples. Previously, individual SVM models have been used for colon classification, however, their performance is limited. In this research study, we propose an SVM-ensemble based new approach for gene based classification of colon, wherein the individual SVM models are constructed through the learning of different SVM kernels, like, linear, polynomial, radial basis function (RBF), and sigmoid. The predicted results of individual models are combined through majority voting. In this way, the combined decision space becomes more discriminative. The proposed technique has been tested on four colon, and several other binary-class gene expression data sets, and improved performance has been achieved compared to previously reported gene based colon cancer detection techniques. The computational time required for the training and testing of 208 × 5,851 data set has been 591.01 and 0.019 s, respectively.

Segmentation and classification of low-quality and noisy ultrasound images is challenging task. In this paper, a new approach is proposed for robust segmentation and classification of carotid artery ultrasound images and consequently, detecting cerebrovascular disease. The proposed technique consists of two phases, in first phase; it refines the class labels selected by user using expectation maximization algorithm. Genetic algorithm is then employed to select discriminative features based on moments of gray-level histogram. The selected features and refined targets are fed as input to neuro-fuzzy classifier for performing segmentation. Finally, intima-media thickness values are measured from segmented images to segregate the normal and abnormal subjects. In second phase, an intelligent decision-making system based on support vector machine is developed to utilize the intima-media thickness values for detecting cerebrovascular disease. The proposed robust segmentation and classification technique for ultrasound images (RSC-US) has been tested on a dataset of 300 real carotid artery ultrasound images and yields accuracy, F-measure, and MCC scores of 98.84, 0.988, 0.9767 %, respectively, using jackknife test. The segmentation and classification performance of the proposed (RSC-US) has been also tested at several noise levels and may be used as secondary observation.

This paper describes the combination of support vector machine (SVM) classifiers using Genetic Programming (GP) for gender classification problem. In our scheme, individual SVM classifiers are constructed through the learning of different SVM kernel functions. The predictions of SVM classifiers are then combined using GP to develop Optimal Composite Classifier (OCC). In this way, the combined decision space is more informative and discriminant. OCC has shown improved performance than that of optimized individual SVM classifiers using grid search. Another advantage of our GP combination scheme is that it automatically incorporates the issues of optimal kernel function and model selection to achieve high performance classification model. The classification performance is reported by using Receiver Operating Characteristics (ROC) Curve. Experiments are conducted under various feature sets to show that OCC is more informative and robust as compared to their individual SVM classifiers. Specifically, it attains high margin of improvement for small feature sets.

In this paper, we have investigated the problem of gender classification using frontal facial images. Four different classifiers, namely K-means, k-nearest neighbors, Linear Discriminant Analysis and Mahalanobis Distance Based classifiers are compared. Receiver operating characteristics (ROC) curve along with the area under the convex hull (AUCH) have been utilized as the performance measures of the classifiers at different feature subsets. To measure the overall performance of a classifier with single scalar value, the new scheme of finding the area under the convex hull of AUCH of ROC curves (AUCH of AUCHS) is proposed. It has been observed that, when the number of macro features is increased beyond 5, the AUCH saturates and even decreases for some classifiers, illustrating the curse of dimensionality. We then used genetic programming to combine classifiers and thus evolved an optimum combined classifier (OCC), producing better performance than the individual classifiers. We found that using only two features, the OCC has comparable performance to that of original classifier using 20 macro features. It produces true positive rate values as high as 0.94 corresponding to false positive rate as low as 0.15 for 1: 3 train to testing ratio. We also observed that heterogeneous combination of classifiers is more promising than the homogenous combination.

In this paper, we consider the challenging problem of anomalous behavior detection in videos. Considering the pixel based anomaly detection, we have used k-mean clustering, a posteriori probability based probabilistic model, and region intersection to detect the anomalies in the video. The proposed technique considers the normal events as the events of higher probabilities. Densely sampled points are passed to a probabilistic model through k-mean clustering to obtain the probability of events. A threshold on the probability values is applied to distinguish anomaly from normal events. The final results of anomalous event detection obtained from different spatial scales are combined through region intersection. The integration of results of multi-scale anomaly detection using region intersection reduces false positives robustly. We have tested our technique over publically available standard video anomaly datase.

In this work we propose a hierarchal feature extraction and reduction framework for object recognition. This framework is motivated by both primate visual cortex and ensemble classification based techniques for object recognition. By understanding the representation of the objects from edges based features, we select PHOG (pyramid histogram of oriented gradients) to describe an object. The curse of PHOG feature dimensionality has been addressed by using a technique which not only reduces dimensions of features, but this feature reduction technique also follows ensemble classification and the hierarchal theory of object recognition in ventral stream of primate visual cortex. Technique follows the feed forward models of object recognition in cortex but is less computationally expensive as compared to other cortex-like mechanisms in literature. We demonstrate the performance of proposed technique over Caltech, MIT-CBCL and INRIA datasets.

In this paper, we proposed a new approach for image clustering to address the adverse effects of noise presented in the images. In particular, the concept of information gain has been incorporated into classical fuzzy c-means (FCM) algorithm in order to develop a robust clustering method. FCM is associated with high sensitivity to noise and produces non-homogenous clustering. To induce robustness to noise, the new clustering technique updates fuzzy membership values and cluster centroids based on information gain. The proposed method produces more homogeneous clustering and its performance can be verified at noisy and noise free images. Experiments have been performed on synthetic, CT liver images and compared with those of classical FCM and one of its robust variants. Moreover, the proposed algorithm has been validated on a data set of 30 carotid artery ultrasound images. Visual inspection of segmented images and clustering quality measures confirm that the proposed approach outperforms other clustering algorithms in comparison. Quantitative measures, in terms of PC and CE, also lead to similar conclusion. Hence, the proposed algorithm is robust to noise and produces homogenous clustering.

Disease diagnostics based on medical imaging is getting popularity day after day. Presence of the arthrosclerosis is one of the causes of narrowing of carotid arteries which may block partially or fully blood flow into the brain. Serious brain strokes may occur due to such types of blockages in blood flow. Early detection of the plaque and taking precautionary steps in this regard may prevent from such type of serious strokes. In this paper, we present automatic image segmentation and decision making technique for carotid artery ultrasound images based on active contour approach. We have successfully applied the automatic segmentation of carotid artery ultrasound images using snake based model. Intima-media thickness (IMT) measurement is used to form a feature vector for classification. Five different features are extracted from IMT measured values. K-nearest neighbors (KNN) classifier is applied for classification of the images. Qualitative comparison of the proposed approach has been made with the manual initialization of snakes for carotid artery image segmentation. Decision is made based on the feature vector obtained from IMT values. Using the proposed approach we have obtained 98.30% classification accuracy. Our proposed approach successfully segment and classify the carotid artery images in an automated way to help radiologists. Obtained results show the effectiveness of the proposed approach.

The current study presents the use of Raman spectroscopy combined with support vector machine (SVM) for the classification of dengue suspected human blood sera. Raman spectra for 84 clinically dengue suspected patients acquired from Holy Family Hospital, Rawalpindi, Pakistan, have been used in this study.The spectral differences between dengue positive and normal sera have been exploited by using effective machine learning techniques. In this regard, SVM models built on the basis of three different kernel functions including Gaussian radial basis function (RBF), polynomial function and linear functionhave been employed to classify the human blood sera based on features obtained from Raman Spectra.The classification model have been evaluated with the 10-fold cross validation method. In the present study, the best performance has been achieved for the polynomial kernel of order 1. A diagnostic accuracy of about 85% with the precision of 90%, sensitivity of 73% and specificity of 93% has been achieved under these conditions.

Image segmentation plays vital role in computer vision, pattern recognition, disease diagnosing in medical images and many other fields. Medical image segmentation is one of the most attractive areas of research, because successfulness of the post-processing techniques highly depends on the accuracy of image segmentation. Further, during acquisition medical images may have been corrupted from various types of degradations and noises. In medical imagery, ultrasound images are usually of poor quality due to the presence of speckle noise and wave interferences. It takes considerable efforts from experts to analyze and make decision on the basis of acquired images. In this paper, we propose an approach called improved spatial fuzzy c-means modified (sFCMM) for image segmentation. This technique works well compared to conventional fuzzy c-means (FCM) and spatial fuzzy c-means (sFCM) algorithms. Three types of features: moments of gray level histogram (MGH), 2D continuous wavelet transforms (2D-CWT) and gray level co-occurrence matrix (GLCM), are extracted from the ultrasound images for segmentation. Clustering validity indices are computed and have been compared with FCM and sFCM algorithm. Experimental results show the effectiveness of the proposed technique.

In this study, novel image clustering algorithm is investigated to improve the clustering performance. We have investigated this model and have achieved improved clustering performance by fine tuning the related model parameters. Yi Yang (2010) proposed clustering algorithm namely local discriminant model and global integration (LDMGI). Clustering parameters are number of nearest neighbours (k) and regularization parameter (λ). The reported parameters are k = 5 and the optimal value of λ selected from set {10−8 – 108} with step size of 102. It is observed that LDMGI clustering performance can be improved with different combination of k and λ. But no criteria exist for the selection of optimal k and λ for best clustering performance. We developed Improved-LDMGI by fine tuning the optimal value of λ in small step size of 0.25 while keeping k = 5 for all image dataset except handwritten image dataset. Significant performance improvement, on average of 7.0 percent, is observed.

It is well known that parameter optimization and dimension reduction strategies play an important role in improving the performance of clustering algorithms. Recently, Yang et at. (2011) proposed an interesting and effective local discriminant model and global integration (LDMGI) clustering algorithm for image databases. We observed this LDMGI approach suffers from the curse of dimensionality. We then show that the effectiveness of this approach could be substantially enhanced with parameter selection and dimensionality reduction approach. We thus experimentally observed the enhanced performance of LDMGI algorithm in terms of clustering accuracy (ACC) and normalized mutual information (NMI). In the first stage, the optimal values of the clustering parameters, nearest neighbours (k) and regularization parameter (λ) are computed. In the second stage, minimum Redundancy and maximum-Relevance (mRMR) technique is utilized to select discriminant image features. During simulation, mRMR based LDMGI model have given an overall 18.7% (ACC) and 11.8% (NMI) higher performance than LDMGI model.

Motivation: Subcellular localization of proteins is one of the most significant characteristics of living cells. Prediction of protein subcellular locations is crucial to the understanding of various protein functions. Therefore, an accurate, computationally efficient and reliable prediction system is required. Results: In this article, the predictions of various Support Vector Machine (SVM) models have been combined through majority voting. The proposed ensemble SVM-SubLoc has achieved the highest success rates of 99.7% using hybrid features of Haralick textures and local binary patterns (HarLBP), 99.4% using hybrid features of Haralick textures and Local Ternary Patterns (HarLTP). In addition, SVM-SubLoc has yielded 99.0% accuracy using only local ternary patterns (LTPs) based features. The dimensionality of HarLBP feature vector is 581 compared with 78 and 52 for HarLTP and LTPs, respectively. Hence, SVM-SubLoc in conjunction with LTPs is fast, sufficiently accurate and simple predictive system. The proposed SVM-SubLoc approach thus provides superior prediction performance using the reduced feature space compared with existing approaches. Availability: A web server accompanying the proposed prediction scheme is available at http://111.68.99.218/SVM-SubLoc Contact:asif@pieas.edu.pk; khan.asifullah@gmail.com Supplementary information:Supplementary data are available at Bioinformatics online.

Membrane proteins are a major class of proteins and encoded by approximately 20% to 30% of genes in most organisms. In this work, a two-layer novel membrane protein prediction system, called Mem-PHybrid, is proposed. It is able to first identify the protein query as a membrane or nonmembrane protein. In the second level, it further identifies the type of membrane protein. The proposed Mem-PHybrid prediction system is based on hybrid features, whereby a fusion of both the physicochemical and split amino acid composition-based features is performed. This enables the proposed Mem-PHybrid to exploit the discrimination capabilities of both types of feature extraction strategy. In addition, minimum redundancy and maximum relevance has also been applied to reduce the dimensionality of a feature vector. We employ random forest, evidence-theoretic K-nearest neighbor, and support vector machine (SVM) as classifiers and analyze their performance on two datasets. SVM using hybrid features yields the highest accuracy of 89.6% and 97.3% on dataset1 and 91.5% and 95.5% on dataset2 for jackknife and independent dataset tests, respectively. The enhanced prediction performance of Mem-PHybrid is largely attributed to the exploitation of the discrimination power of the hybrid features and of the learning capability of SVM. Mem-PHybrid is accessible at http://www.111.68.99.218/Mem-PHybrid.

Knowledge of the types of membrane protein provides useful clues in deducing the functions of uncharacterized membrane proteins. An automatic method for efficiently identifying uncharacterized proteins is thus highly desirable. In this work, we have developed a novel method for predicting membrane protein types by exploiting the discrimination capability of the difference in amino acid composition at the N and C terminus through split amino acid composition (SAAC). We also show that the ensemble classification can better exploit this discriminating capability of SAAC. In this study, membrane protein types are classified using three feature extraction and several classification strategies. An ensemble classifier Mem-EnsSAAC is then developed using the best feature extraction strategy. Pseudo amino acid (PseAA) composition, discrete wavelet analysis (DWT), SAAC, and a hybrid model are employed for feature extraction. The nearest neighbor, probabilistic neural network, support vector machine, random forest, and Adaboost are used as individual classifiers. The predicted results of the individual learners are combined using genetic algorithm to form an ensemble classifier, Mem-EnsSAAC yielding an accuracy of 92.4 and 92.2% for the Jackknife and independent dataset test, respectively. Performance measures such as MCC, sensitivity, specificity, F-measure, and Q-statistics show that SAAC-based prediction yields significantly higher performance compared to PseAA- and DWT-based systems, and is also the best reported so far. The proposed Mem-EnsSAAC is able to predict the membrane protein types with high accuracy and consequently, can be very helpful in drug discovery. It can be accessed at http://111.68.99.218/membrane.

Precise information about protein locations in a cell facilitates in the understanding of the function of a protein and its interaction in the cellular environment. This information further helps in the study of the specific metabolic pathways and other biological processes. We propose an ensemble approach called ''CE-PLoc'' for predicting subcellular locations based on fusion of individual classifiers. The proposed approach utilizes features obtained from both dipeptide composition (DC) and amphiphilic pseudo amino acid composition (PseAAC) based feature extraction strategies. Different feature spaces are obtained by varying the dimensionality using PseAAC for a selected base learner. The performance of the individual learning mechanisms such as support vector machine, nearest neighbor, probabilistic neural network, covariant discriminant, which are trained using PseAAC based features is first analyzed. Classifiers are developed using same learning mechanism but trained on PseAAC based feature spaces of varying dimensions. These classifiers are combined through voting strategy and an improvement in prediction performance is achieved. Prediction performance is further enhanced by developing CE-PLoc through the combination of different learning mechanisms trained on both DC based feature space and PseAAC based feature spaces of varying dimensions. The predictive performance of proposed CE-PLoc is evaluated for two benchmark datasets of protein subcellular locations using accuracy, MCC, and Q-statistics. Using the jackknife test, prediction accuracies of 81.47 and 83.99% are obtained for 12 and 14 subcellular locations datasets, respectively. In case of independent dataset test, prediction accuracies are 87.04 and 87.33% for 12 and 14 class datasets, respectively.

Membrane proteins play an important role in many biological processes and are attractive drug targets. In this study, membrane proteins are classified using two feature extraction and several classification strategies. The first feature extraction strategy is pseudo amino acid (PseAA) composition; utilizing hydrophobicity and hydrophilicity for reflecting the sequence order effects, while the second method is discrete wavelet analysis (DWT); analyzing the different components of a signal localized both in space and scale domains. The nearest neighbor, probabilistic neural network, support vector machine, random forest, and Adaboost are used as basic learning mechanisms. The predicted results of the base learners are combined using majority voting to form an ensemble classifier. The best accuracy obtained for the Jackknife and independent dataset test is 85.4% and 95.3%, respectively. Using performance measures such as MCC, Sensitivity, Specificity, and F-measure, it has been observed that PseAA based prediction is significantly higher than that of the DWT, and is also the best reported, so far.

Membrane proteins are vital type of proteins that serve as channels, receptors, and energy transducers in a cell. Prediction of membrane protein types is an important research area in bioinformatics. Knowledge of membrane protein types provides some valuable information for predicting novel example of the membrane protein types. However, classification of membrane protein types can be both time consuming and susceptible to errors due to the inherent similarity of membrane protein types. In this paper, neural networks based membrane protein type prediction system is proposed. Composite protein sequence representation (CPSR) is used to extract the features of a protein sequence, which includes seven feature sets; amino acid composition, sequence length, 2 gram exchange group frequency, hydrophobic group, electronic group, sum of hydrophobicity, and R-group. Principal component analysis is then employed to reduce the dimensionality of the feature vector. The probabilistic neural network (PNN), generalized regression neural network, and support vector machine (SVM) are used as classifiers. A high success rate of 86.01% is obtained using SVM for the jackknife test. In case of independent dataset test, PNN yields the highest accuracy of 95.73%. These classifiers exhibit improved performance using other performance measures such as sensitivity, specificity, Mathew's correlation coefficient, and F-measure. The experimental results show that the prediction performance of the proposed scheme for classifying membrane protein types is the best reported, so far. This performance improvement may largely be credited to the learning capabilities of neural networks and the composite feature extraction strategy, which exploits seven different properties of protein sequences. The proposed Mem-Predictor can be accessed at http://111.68.99.218/Mem-Predictor.