Synergic Deep Learning for Smart Health Diagnosis of COVID-19 for Connected Living and Smart Cities

COVID-19 pandemic has led to a significant loss of global deaths, economical status, and so on. To prevent and control COVID-19, a range of smart, complex, spatially heterogeneous, control solutions, and strategies have been conducted. Earlier classification of 2019 novel coronavirus disease (COVID-19) is needed to cure and control the disease. It results in a requirement of secondary diagnosis models, since no precise automated toolkits exist. The latest finding attained using radiological imaging techniques highlighted that the images hold noticeable details regarding the COVID-19 virus. The application of recent artificial intelligence (AI) and deep learning (DL) approaches integrated to radiological images finds useful to accurately detect the disease. This article introduces a new synergic deep learning (SDL)-based smart health diagnosis of COVID-19 using Chest X-Ray Images. The SDL makes use of dual deep convolutional neural networks (DCNNs) and involves a mutual learning process from one another. Particularly, the representation of images learned by both DCNNs is provided as the input of a synergic network, which has a fully connected structure and predicts whether the pair of input images come under the identical class. Besides, the proposed SDL model involves a fuzzy bilateral filtering (FBF) model to pre-process the input image. The integration of FBL and SDL resulted in the effective classification of COVID-19. To investigate the classifier outcome of the SDL model, a detailed set of simulations takes place and ensures the effective performance of the FBF-SDL model over the compared methods.

A Deep Learning Approach for Basal Cell Carcinomas and Bowen’s Disease Recognition in Dermatopathology Image

Basal cell carcinomas and Bowen’s disease (squamous cell carcinoma in situ) are the most common cutaneous tumors. The early diagnoses of these diseases are very important due to their better prognosis. But it is a heavy workload for the pathologists to recognize a large number of pathological images and diagnose these diseases. So, there is an urgent need to develop an automatic method for detecting and classifying the skin cancers. This paper presents a recognition system of dermatopathology images based on the deep convolutional neural networks (CNN). The dermatopathology images are collected from the hospital. The deep learning model is trained using different image datasets. It can be found that the recognition accuracy of the system can be improved by using data augmentation even if the number of the clinical samples are not increased. But the recognition accuracy of the system is the highest when the number of the original histological image is increased. The experimental results that the system can correctly recognize 88.5% of patients with basal cell carcinoma and 86.5% of patients with Bowen’s disease.

Earth Imagery Segmentation on Terrain Surface with Limited Training Labels: A Semi-supervised Approach based on Physics-Guided Graph Co-Training

Given earth imagery with spectral features on a terrain surface, this paper studies surface segmentation based on both explanatory features and surface topology. The problem is important in many spatial and spatiotemporal applications such as flood extent mapping in hydrology. The problem is uniquely challenging for several reasons: first, the size of earth imagery on a terrain surface is often much larger than the input of popular deep convolutional neural networks; second, there exists topological structure dependency between pixel classes on the surface, and such dependency can follow an unknown and non-linear distribution; third, there are often limited training labels. Existing methods for earth imagery segmentation often divide the imagery into patches and consider the elevation as an additional feature channel. These methods do not fully incorporate the spatial topological structural constraint within and across surface patches and thus often show poor results, especially when training labels are limited. Existing methods on semi-supervised and unsupervised learning for earth imagery often focus on learning representation without explicitly incorporating surface topology. In contrast, we propose a novel framework that explicitly models the topological skeleton of a terrain surface with a contour tree from computational topology, which is guided by the physical constraint (e.g., water flow direction on terrains). Our framework consists of two neural networks: a convolutional neural network (CNN) to learn spatial contextual features on a 2D image grid, and a graph neural network (GNN) to learn the statistical distribution of physics-guided spatial topological dependency on the contour tree. The two models are co-trained via variational EM. Evaluations on the real-world flood mapping datasets show that the proposed models outperform baseline methods in classification accuracy, especially when training labels are limited.

xDNN: Inference for Deep Convolutional Neural Networks

We present xDNN, an end-to-end system for deep-learning inference based on a family of specialized hardware processors synthesized on Field-Programmable Gate Array (FPGAs) and Convolution Neural Networks (CNN). We present a design optimized for low latency, high throughput, and high compute efficiency with no batching. The design is scalable and a parametric function of the number of multiply-accumulate units, on-chip memory hierarchy, and numerical precision. The design can produce a scale-down processor for embedded devices, replicated to produce more cores for larger devices, or resized to optimize efficiency. On Xilinx Virtex Ultrascale+ VU13P FPGA, we achieve 800 MHz that is close to the Digital Signal Processing maximum frequency and above 80% efficiency of on-chip compute resources. On top of our processor family, we present a runtime system enabling the execution of different networks for different input sizes (i.e., from 224× 224 to 2048× 1024). We present a compiler that reads CNNs from native frameworks (i.e., MXNet, Caffe, Keras, and Tensorflow), optimizes them, generates codes, and provides performance estimates. The compiler combines quantization information from the native environment and optimizations to feed the runtime with code as efficient as any hardware expert could write. We present tools partitioning a CNN into subgraphs for the division of work to CPU cores and FPGAs. Notice that the software will not change when or if the FPGA design becomes an ASIC, making our work vertical and not just a proof-of-concept FPGA project. We show experimental results for accuracy, latency, and power for several networks: In summary, we can achieve up to 4 times higher throughput, 3 times better power efficiency than the GPUs, and up to 20 times higher throughput than the latest CPUs. To our knowledge, we provide solutions faster than any previous FPGA-based solutions and comparable to any other top-of-the-shelves solutions.

Automatic Parotid Gland Segmentation in MVCT Using Deep Convolutional Neural Networks

Radiation-induced xerostomia, as a major problem in radiation treatment of the head and neck cancer, is mainly due to the overdose irradiation injury to the parotid glands. Helical Tomotherapy-based megavoltage computed tomography (MVCT) imaging during the Tomotherapy treatment can be applied to monitor the successive variations in the parotid glands. While manual segmentation is time consuming, laborious, and subjective, automatic segmentation is quite challenging due to the complicated anatomical environment of head and neck as well as noises in MVCT images. In this article, we propose a localization-refinement scheme to segment the parotid gland in MVCT. After data pre-processing we use mask region convolutional neural network (Mask R-CNN) in the localization stage after data pre-processing, and design a modified U-Net in the following fine segmentation stage. To the best of our knowledge, this study is a pioneering work of deep learning on MVCT segmentation. Comprehensive experiments based on different data distribution of head and neck MVCTs and different segmentation models have demonstrated the superiority of our approach in terms of accuracy, effectiveness, flexibility, and practicability. Our method can be adopted as a powerful tool for radiation-induced injury studies, where accurate organ segmentation is crucial.

Deep segmentation of the liver and the hepatic tumors from abdomen tomography images

A pipelined framework is proposed for accurate, automated, simultaneous segmentation of the liver as well as the hepatic tumors from computed tomography (CT) images. The introduced framework composed of three pipelined levels. First, two different transfers deep convolutional neural networks (CNN) are applied to get high-level compact features of CT images. Second, a pixel-wise classifier is used to obtain two output-classified maps for each CNN model. Finally, a fusion neural network (FNN) is used to integrate the two maps. Experimentations performed on the MICCAI’2017 database of the liver tumor segmentation (LITS) challenge, result in a dice similarity coefficient (DSC) of 93.5% for the segmentation of the liver and of 74.40% for the segmentation of the lesion, using a 5-fold cross-validation scheme. Comparative results with the state-of-the-art techniques on the same data show the competing performance of the proposed framework for simultaneous liver and tumor segmentation.