Shawn Xiang Li

Bio: Shawn Xiang Li is an academic researcher from Shenzhen University. The author has contributed to research in topics: Segmentation & Deep learning. The author has an hindex of 2, co-authored 2 publications receiving 248 citations.

The bumps under the hippocampus.

Abstract: Shown in every neuroanatomy textbook, a key morphological feature is the bumpy ridges, which we refer to as hippocampal dentation, on the inferior aspect of the hippocampus. Like the folding of the cerebral cortex, hippocampal dentation allows for greater surface area in a confined space. However, examining numerous approaches to hippocampal segmentation and morphology analysis, virtually all published 3D renderings of the hippocampus show the inferior surface to be quite smooth or mildly irregular; we have rarely seen the characteristic bumpy structure on reconstructed 3D surfaces. The only exception is a 9.4T postmortem study (Yushkevich et al. [2009]: NeuroImage 44:385-398). An apparent question is, does this indicate that this specific morphological signature can only be captured using ultra high-resolution techniques? Or, is such information buried in the data we commonly acquire, awaiting a computation technique that can extract and render it clearly? In this study, we propose an automatic and robust super-resolution technique that captures the fine scale morphometric features of the hippocampus based on common 3T MR images. The method is validated on 9.4T ultra-high field images and then applied on 3T data sets. This method opens possibilities of future research on the hippocampus and other sub-cortical structural morphometry correlating the degree of dentation with a range of diseases including epilepsy, Alzheimer's disease, and schizophrenia. Hum Brain Mapp 39:472-490, 2018. © 2017 Wiley Periodicals, Inc.

An end-to-end multi-task system of automatic lesion detection and anatomical localization in whole-body bone scintigraphy by deep learning

Abstract: Abstract Summary Limited by spatial resolution and visual contrast, bone scintigraphy interpretation is susceptible to subjective factors, which considerably affects the accuracy and repeatability of lesion detection and anatomical localization. In this work, we design and implement an end-to-end multi-task deep learning model to perform automatic lesion detection and anatomical localization in whole-body bone scintigraphy. A total of 617 whole-body bone scintigraphy cases including anterior and posterior views were retrospectively analyzed. The proposed semi-supervised model consists of two task flows. The first one, the lesion segmentation flow, received image patches and was trained in a supervised way. The other one, skeleton segmentation flow, was trained on as few as five labeled images in conjunction with the multi-atlas approach, in a semi-supervised way. The two flows joint in their encoder layers so each flow can capture more generalized distribution of the sample space and extract more abstract deep features. The experimental results show that the architecture achieved the highest precision in the finest bone segmentation task in both anterior and posterior images of whole-body scintigraphy. Such an end-to-end approach with very few manual annotation requirement would be suitable for algorithm deployment. Moreover, the proposed approach reliably balances unsupervised labels construction and supervised learning, providing useful insight for weakly labeled image analysis. Supplementary information Supplementary data are available at Bioinformatics online.

Open-source algorithm and software for computed tomography-based virtual pancreatoscopy and other applications

Abstract: Abstract Pancreatoscopy plays a significant role in the diagnosis and treatment of pancreatic diseases. However, the risk of pancreatoscopy is remarkably greater than that of other endoscopic procedures, such as gastroscopy and bronchoscopy, owing to its severe invasiveness. In comparison, virtual pancreatoscopy (VP) has shown notable advantages. However, because of the low resolution of current computed tomography (CT) technology and the small diameter of the pancreatic duct, VP has limited clinical use. In this study, an optimal path algorithm and super-resolution technique are investigated for the development of an open-source software platform for VP based on 3D Slicer. The proposed segmentation of the pancreatic duct from the abdominal CT images reached an average Dice coefficient of 0.85 with a standard deviation of 0.04. Owing to the excellent segmentation performance, a fly-through visualization of both the inside and outside of the duct was successfully reconstructed, thereby demonstrating the feasibility of VP. In addition, a quantitative analysis of the wall thickness and topology of the duct provides more insight into pancreatic diseases than a fly-through visualization. The entire VP system developed in this study is available at https://github.com/gaoyi/VirtualEndoscopy.git .

Open-source algorithm and software for computed tomography-based virtual pancreatoscopy and other applications

Abstract: Abstract Pancreatoscopy plays a significant role in the diagnosis and treatment of pancreatic diseases. However, the risk of pancreatoscopy is remarkably greater than that of other endoscopic procedures, such as gastroscopy and bronchoscopy, owing to its severe invasiveness. In comparison, virtual pancreatoscopy (VP) has shown notable advantages. However, because of the low resolution of current computed tomography (CT) technology and the small diameter of the pancreatic duct, VP has limited clinical use. In this study, an optimal path algorithm and super-resolution technique are investigated for the development of an open-source software platform for VP based on 3D Slicer. The proposed segmentation of the pancreatic duct from the abdominal CT images reached an average Dice coefficient of 0.85 with a standard deviation of 0.04. Owing to the excellent segmentation performance, a fly-through visualization of both the inside and outside of the duct was successfully reconstructed, thereby demonstrating the feasibility of VP. In addition, a quantitative analysis of the wall thickness and topology of the duct provides more insight into pancreatic diseases than a fly-through visualization. The entire VP system developed in this study is available at https://github.com/gaoyi/VirtualEndoscopy.git .

Deep Learning for Classification and Localization of COVID-19 Markers in Point-of-Care Lung Ultrasound

Abstract: Deep learning (DL) has proved successful in medical imaging and, in the wake of the recent COVID-19 pandemic, some works have started to investigate DL-based solutions for the assisted diagnosis of lung diseases. While existing works focus on CT scans, this paper studies the application of DL techniques for the analysis of lung ultrasonography (LUS) images. Specifically, we present a novel fully-annotated dataset of LUS images collected from several Italian hospitals, with labels indicating the degree of disease severity at a frame-level, video-level, and pixel-level (segmentation masks). Leveraging these data, we introduce several deep models that address relevant tasks for the automatic analysis of LUS images. In particular, we present a novel deep network, derived from Spatial Transformer Networks, which simultaneously predicts the disease severity score associated to a input frame and provides localization of pathological artefacts in a weakly-supervised way. Furthermore, we introduce a new method based on uninorms for effective frame score aggregation at a video-level. Finally, we benchmark state of the art deep models for estimating pixel-level segmentations of COVID-19 imaging biomarkers. Experiments on the proposed dataset demonstrate satisfactory results on all the considered tasks, paving the way to future research on DL for the assisted diagnosis of COVID-19 from LUS data.

A review of deep learning with special emphasis on architectures, applications and recent trends

Abstract: Deep learning (DL) has solved a problem that a few years ago was thought to be intractable — the automatic recognition of patterns in spatial and temporal data with an accuracy superior to that of humans. It has solved problems beyond the realm of traditional, hand-crafted machine learning algorithms and captured the imagination of practitioners who are inundated with all types of data. As public awareness of the efficacy of DL increases so does the desire to make use of it. But even for highly trained professionals it can be daunting to approach the rapidly increasing body of knowledge in the field. Where does one start? How does one determine if a particular DL model is applicable to their problem? How does one train and deploy them? With these questions in mind, we present an overview of some of the key DL architectures. We also discuss some new automatic architecture optimization protocols that use multi-agent approaches. Further, since guaranteeing system uptime is critical to many applications, a section dwells on using DL for fault detection and mitigation. This is followed by an exploratory survey of several areas where DL emerged as a game-changer: fraud detection in financial applications, financial time-series forecasting, predictive and prescriptive analytics, medical image processing, power systems research and recommender systems. The thrust of this review is to outline emerging applications of DL and provide a reference to researchers seeking to use DL in their work for pattern recognition with unparalleled learning capacity and the ability to scale with data.

Deep Learning in Ultrasound Imaging

Abstract: In this article, we consider deep learning strategies in ultrasound systems, from the front end to advanced applications. Our goal is to provide the reader with a broad understanding of the possible impact of deep learning methodologies on many aspects of ultrasound imaging. In particular, we discuss methods that lie at the interface of signal acquisition and machine learning, exploiting both data structure (e.g., sparsity in some domain) and data dimensionality (big data) already at the raw radio-frequency channel stage. As some examples, we outline efficient and effective deep learning solutions for adaptive beamforming and adaptive spectral Doppler through artificial agents, learn compressive encodings for the color Doppler, and provide a framework for structured signal recovery by learning fast approximations of iterative minimization problems, with applications to clutter suppression and super-resolution ultrasound. These emerging technologies may have a considerable impact on ultrasound imaging, showing promise across key components in the receive processing chain.

A new method for prediction of air pollution based on intelligent computation

Abstract: The detection and treatment of increasing air pollution due to technological developments represent some of the most important challenges facing the world today. Indeed, there has been a significant increase in levels of environmental pollution in recent years. The aim of the work presented herein is to design an intelligent predictor for the concentrations of air pollutants over the next 2 days based on deep learning techniques using a recurrent neural network (RNN). The best structure for its operation is then determined using a particle swarm optimization (PSO) algorithm. The new predictor based on intelligent computation relying on unsupervised learning, i.e., long short-term memory (LSTM) and optimization (i.e., PSO), is called the smart air quality prediction model (SAQPM). The main goal is to predict six the concentrations of six types of air pollution, viz. PM2.5 particulate matter, PM10, particulate matter, nitrogen dioxide (NO2), carbon monoxide (CO), ozone (O3), and sulfur dioxide (SO2). SAQPM consists of four stages. The first stage involves data collection from multiple stations (35 in this case). The second stage involves preprocessing of the data, including (a) separation of each station with an independent focus, (b) handle missing values, and (c) normalization of the dataset to the range of (0, 1) using the MinMaxScalar method. The third stage relates to building the predictor based on the LSTM method by identifying the best structure and parameter values (weight, bias, number of hidden layers, number of nodes in each hidden layer, and activation function) for the network using the functional PSO algorithm to achieve a goal. Thereafter, the dataset is split into training and testing parts based on the ten cross-validation principle. The training dataset is then used to build the predictor. In the fourth stage, evaluation results for each station are obtained by reading the concentration of each pollutant each hour for at most 30 days then taking the average of the symmetric mean absolute percentage error (SMAPE) for 25 days only.

Deep Learning to Obtain Simultaneous Image and Segmentation Outputs From a Single Input of Raw Ultrasound Channel Data

Abstract: Single plane wave transmissions are promising for automated imaging tasks requiring high ultrasound frame rates over an extended field of view. However, a single plane wave insonification typically produces suboptimal image quality. To address this limitation, we are exploring the use of deep neural networks (DNNs) as an alternative to delay-and-sum (DAS) beamforming. The objectives of this work are to obtain information directly from raw channel data and to simultaneously generate both a segmentation map for automated ultrasound tasks and a corresponding ultrasound B-mode image for interpretable supervision of the automation. We focus on visualizing and segmenting anechoic targets surrounded by tissue and ignoring or deemphasizing less important surrounding structures. DNNs trained with Field II simulations were tested with simulated, experimental phantom, and in vivo data sets that were not included during training. With unfocused input channel data (i.e., prior to the application of receive time delays), simulated, experimental phantom, and in vivo test data sets achieved mean ± standard deviation Dice similarity coefficients of 0.92 ± 0.13, 0.92 ± 0.03, and 0.77 ± 0.07, respectively, and generalized contrast-to-noise ratios (gCNRs) of 0.95 ± 0.08, 0.93 ± 0.08, and 0.75 ± 0.14, respectively. With subaperture beamformed channel data and a modification to the input layer of the DNN architecture to accept these data, the fidelity of image reconstruction increased (e.g., mean gCNR of multiple acquisitions of two in vivo breast cysts ranged 0.89–0.96), but DNN display frame rates were reduced from 395 to 287 Hz. Overall, the DNNs successfully translated feature representations learned from simulated data to phantom and in vivo data, which is promising for this novel approach to simultaneous ultrasound image formation and segmentation.