Showing papers by "Asifullah Khan published in 2020"

A survey of the recent architectures of deep convolutional neural networks

Abstract: Deep Convolutional Neural Network (CNN) is a special type of Neural Networks, which has shown exemplary performance on several competitions related to Computer Vision and Image Processing. Some of the exciting application areas of CNN include Image Classification and Segmentation, Object Detection, Video Processing, Natural Language Processing, and Speech Recognition. The powerful learning ability of deep CNN is primarily due to the use of multiple feature extraction stages that can automatically learn representations from the data. The availability of a large amount of data and improvement in the hardware technology has accelerated the research in CNNs, and recently interesting deep CNN architectures have been reported. Several inspiring ideas to bring advancements in CNNs have been explored, such as the use of different activation and loss functions, parameter optimization, regularization, and architectural innovations. However, the significant improvement in the representational capacity of the deep CNN is achieved through architectural innovations. Notably, the ideas of exploiting spatial and channel information, depth and width of architecture, and multi-path information processing have gained substantial attention. Similarly, the idea of using a block of layers as a structural unit is also gaining popularity. This survey thus focuses on the intrinsic taxonomy present in the recently reported deep CNN architectures and, consequently, classifies the recent innovations in CNN architectures into seven different categories. These seven categories are based on spatial exploitation, depth, multi-path, width, feature-map exploitation, channel boosting, and attention. Additionally, the elementary understanding of CNN components, current challenges, and applications of CNN are also provided.

Intrusion detection using deep sparse auto-encoder and self-taught learning

Abstract: 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.

COVID-19 Detection in Chest X-Ray Images using a New Channel Boosted CNN

Abstract: COVID-19 is a highly contagious respiratory infection that has affected a large population across the world and continues with its devastating consequences. It is imperative to detect COVID-19 at the earliest to limit the span of infection. In this work, a new classification technique CB-STM-RENet based on deep Convolutional Neural Network (CNN) and Channel Boosting is proposed for the screening of COVID-19 in chest X-Rays. In this connection, to learn the COVID-19 specific radiographic patterns, a new convolution block based on split-transform-merge (STM) is developed. This new block systematically incorporates region and edge-based operations at each branch to capture the diverse set of features at various levels, especially those related to region homogeneity, textural variations, and boundaries of the infected region. The learning and discrimination capability of the proposed CNN architecture is enhanced by exploiting the Channel Boosting idea that concatenates the auxiliary channels along with the original channels. The auxiliary channels are generated from the pre-trained CNNs using Transfer Learning. The effectiveness of the proposed technique CB-STM-RENet is evaluated on three different datasets of chest X-Rays namely CoV-Healthy-6k, CoV-NonCoV-10k, and CoV-NonCoV-15k. The performance comparison of the proposed CB-STM-RENet with the existing techniques exhibits high performance both in discriminating COVID-19 chest infections from Healthy, as well as, other types of chest infections. CB-STM-RENet provides the highest performance on all these three datasets; especially on the stringent CoV-NonCoV-15k dataset. The good detection rate (97%), and high precision (93%) of the proposed technique suggest that it can be adapted for the diagnosis of COVID-19 infected patients. The test code is available at this https URL.

Classification and Region Analysis of COVID-19 Infection using Lung CT Images and Deep Convolutional Neural Networks

Abstract: COVID-19 is a global health problem. Consequently, early detection and analysis of the infection patterns are crucial for controlling infection spread as well as devising a treatment plan. This work proposes a two-stage deep Convolutional Neural Networks (CNNs) based framework for delineation of COVID-19 infected regions in Lung CT images. In the first stage, initially, COVID-19 specific CT image features are enhanced using a two-level discrete wavelet transformation. These enhanced CT images are then classified using the proposed custom-made deep CoV-CTNet. In the second stage, the CT images classified as infectious images are provided to the segmentation models for the identification and analysis of COVID-19 infectious regions. In this regard, we propose a novel semantic segmentation model CoV-RASeg, which systematically uses average and max pooling operations in the encoder and decoder blocks. This systematic utilization of max and average pooling operations helps the proposed CoV-RASeg in simultaneously learning both the boundaries and region homogeneity. Moreover, the idea of attention is incorporated to deal with mildly infected regions. The proposed two-stage framework is evaluated on a standard Lung CT image dataset, and its performance is compared with the existing deep CNN models. The performance of the proposed CoV-CTNet is evaluated using Mathew Correlation Coefficient (MCC) measure (0.98) and that of proposed CoV-RASeg using Dice Similarity (DS) score (0.95). The promising results on an unseen test set suggest that the proposed framework has the potential to help the radiologists in the identification and analysis of COVID-19 infected regions.

Channel Boosted Convolutional Neural Network for Classification of Mitotic Nuclei using Histopathological Images

Abstract: 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.

Deep Object Detection based Mitosis Analysis in Breast Cancer Histopathological Images.

Abstract: Empirical evaluation of breast tissue biopsies for mitotic nuclei detection is considered an important prognostic biomarker in tumor grading and cancer progression. However, automated mitotic nuclei detection poses several challenges because of the unavailability of pixel-level annotations, different morphological configurations of mitotic nuclei, their sparse representation, and close resemblance with non-mitotic nuclei. These challenges undermine the precision of the automated detection model and thus make detection difficult in a single phase. This work proposes an end-to-end detection system for mitotic nuclei identification in breast cancer histopathological images. Deep object detection-based Mask R-CNN is adapted for mitotic nuclei detection that initially selects the candidate mitotic region with maximum recall. However, in the second phase, these candidate regions are refined by multi-object loss function to improve the precision. The performance of the proposed detection model shows improved discrimination ability (F-score of 0.86) for mitotic nuclei with significant precision (0.86) as compared to the two-stage detection models (F-score of 0.701) on TUPAC16 dataset. Promising results suggest that the deep object detection-based model has the potential to learn the characteristic features of mitotic nuclei from weakly annotated data and suggests that it can be adapted for the identification of other nuclear bodies in histopathological images.

Sliding mode control for a fractional-order non-linear glucose-insulin system.

Abstract: By providing the generalisation of integration and differentiation, and incorporating the memory and hereditary effects, fractional-order modelling has gotten significant attention in the past few years. One of the extensively studied and utilised models to describe the glucose-insulin system of a human body is Bergman's minimal model. This non-linear model comprises of integer-order differential equations. However, comparison with the experimental data shows that the fractional-order version of Bergman's minimal model is a better representative of the glucose-insulin system than its original integer-order model. To design a control law for an artificial pancreas for a diabetic patient using a fractional-order model, different techniques, including feedback linearisation, have been applied in the literature. The authors' previous work shows that the fractional-order version of Bergman's model describes the glucose-insulin system in a better way than the integer-order model. This study applies the sliding mode control technique and then compares the obtained simulation results with the ones obtained using feedback linearisation.

Autonomous Navigation of Swarms in 3D Environments Using Deep Reinforcement Learning

Abstract: The presence of swarm intelligence in many natural systems has always been an inspiration to develop such distributed intelligence in artificial multi-agent systems. It finds its applications in high-level control of complex swarms, distributed sensing technologies, and telecom networks. In this paper, we present an end-to-end approach to train a group of cooperative agents to navigate through 3-dimensional (3D) environments. The problem is particularly hard because the agents can only observe the environment partially and the number of agents in the swarm (also known as the size of the swarm) may change over time. Our approach uses deep reinforcement learning, mapping raw sensory data to high-level commands, in order to optimize (1) navigation, and (2) distributed assembly of the swarm while keeping the swarm (3) unaffected from its size dynamics. Here, we use suitable reward shaping for navigation and distributed assembly and deal size dynamics by exploiting histograms as an observational input to the model. The simulations were performed in the Unity 3D engine. The results demonstrate that our approach effectively improves swarm navigation and assembly in rough 3D environments and can be generalized to real-world scenarios.

Extracting Signals of Higgs Boson From Background Noise Using Deep Neural Networks.

Abstract: Higgs boson is a fundamental particle, and the classification of Higgs signals is a well-known problem in high energy physics. The identification of the Higgs signal is a challenging task because its signal has a resemblance to the background signals. This study proposes a Higgs signal classification using a novel combination of random forest, auto encoder and deep auto encoder to build a robust and generalized Higgs boson prediction system to discriminate the Higgs signal from the background noise. The proposed ensemble technique is based on achieving diversity in the decision space, and the results show good discrimination power on the private leaderboard; achieving an area under the Receiver Operating Characteristic curve of 0.9 and an Approximate Median Significance score of 3.429.

Lymphocyte Annotator: CD3 + and CD8 + IHC Stained Patch Image Annotation Tool

Abstract: 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.

Wind Speed Prediction using Deep Ensemble Learning with a Jet-like Architecture

Abstract: The wind is one of the most increasingly used renewable energy resources. Accurate and reliable forecast of wind speed is necessary for efficient power production; however, it is not an easy task because it depends upon meteorological features of the surrounding region. Deep learning is extensively used these days for performing feature extraction. It has also been observed that the integration of several learning models, known as ensemble learning, generally gives better performance compared to a single model. The design of wings, tail, and nose of a jet improves the aerodynamics resulting in a smooth and controlled flight of the jet against the variations of the air currents. Inspired by the shape and working of a jet, a novel Deep Ensemble Learning using Jet-like Architecture (DEL-Jet) technique is proposed to enhance the diversity and robustness of a learning system against the variations in the input space. The diverse feature spaces of the base-regressors are exploited using the jet-like ensemble architecture. Two Convolutional Neural Networks (as jet wings) and one deep Auto-Encoder (as jet tail) are used to extract the diverse feature spaces from the input data. After that, nonlinear PCA (as jet main body) is employed to reduce the dimensionality of extracted feature space. Finally, both the reduced and the original feature spaces are exploited to train the meta-regressor (as jet nose) for forecasting the wind speed. The performance of the proposed DEL-Jet technique is evaluated for ten independent runs and shows that the deep and jet-like architecture helps in improving the robustness and generalization of the learning system.

Identification of Malware Families for Creating Generic Signatures : Using Dynamic Analysis and Clustering Methods *

Abstract: An important step in fighting malware is the creation of generalized signatures for the detection and removal of these malware. Millions of new samples are received in anti-malware research labs every day. Generating signatures for these malware requires many techniques i.e. static and dynamic analysis, reverse engineering and identification of malware families, etc. In this research, we used unsupervised learning methods to identify various malware clusters (families). Once similar malware is clustered together then there is no need for generating unique one-to-one signatures to detect this similar malware. Instead, only a generalized signature is enough to detect most of the malware in a cluster. This approach not only speeds up the detection of malware but also reduces the frequency and volume of signature updates on client-side anti-malware applications. We performed a dynamic analysis of 1247 malware samples of different families in a controlled environment (cuckoo sandbox) and features relevant to classification /clustering were extracted through python scripts (1194 features from each sample). Then machine learning methods for unsupervised learning were trained via these features. K-means, Mini-batch K-means, Agglomerative clustering, spectral clustering, and density-based clustering methods were applied to our dataset and 10 distinct clusters were identified based on best scores. Clustering being a heuristic approach, performed well in this work. Visualization of resulting clusters/groups confirmed the presence of different families. The best score was obtained using K Means and mini-batch K Means with n=10.

Development of a Low-Cost Sensor-Based Adaptive Gripper

Abstract: Grippers are the devices that mimic human hands in both function and form. Most of the grippers; however lack sensing abilities. The ones integrated with feedback and sensors are quite expensive. Safe and reliable grasping also demands a systematic and computationally viable algorithm. This paper presents the development of a low-cost sensor-based adaptive gripper with three fingers controlled by a motor. For efficient grasping of an unknown object, the paper presents a novel algorithm. Experiments show that the developed underactuated gripper is able to grasp irregular objects without damaging them. Sensing abilities, and the application of the proposed algorithm enable the detection and prevention of the slippage of the grasped object.