Navigation bronchoscopy is a minimally invasive procedure in which doctors pass a bronchoscope into a subject’s airways to sample the target pulmonary lesion. A three-dimensional (3D) airway roadmap reconstructed from Computer Tomography (CT) scans is a prerequisite for this procedure, especially when the target is distally located. Therefore, an accurate and efficient airway segmentation algorithm is essential to reduce bronchoscopists’ burden of pre-procedural airway identification as well as patients’ discomfort during the prolonged procedure. However, airway segmentation remains a challenging task because of the intrinsic complex tree-like structure, imbalanced sizes of airway branches, potential domain shifts of CT scans, and few available labeled images. To address these problems, we present a deep learning-based pipeline, denoted as NaviAirway, which finds finer bronchioles through four major novel components – feature extractor modules in model architecture design, a bronchiole-sensitive loss function, a human-vision-inspired iterative training strategy, and a semi-supervised learning framework to utilize unlabeled CT images. Experimental results showed that NaviAirway outperformed existing methods, particularly in identification of higher generation bronchioles and robustness to new CT scans. On average, NaviAirway takes five minutes to segment the CT scans of one patient on a GPU-embedded computer. Moreover, we propose two new metrics to complement conventional ones for a more comprehensive and fairer evaluation of deep learning-based airway segmentation approaches. The code is publicly available on https://github.com/AntonotnaWang/NaviAirway.

With the adoption of edge computing, several data centers are available within the footprint of an optical metro network, and contents are replicated in multiple locations. Such a wide content replication offers a unique opportunity to provide better services to users, especially for content-based services, e.g., video delivery. Thus, a service-provisioning scheme can embrace this opportunity to optimize network resource utilization, improve reliability, and achieve lower latency. In this study, we propose a reliable service-provisioning scheme that selects the optimal subset of data centers hosting the desired content and inversely multiplexes a content request over multiple link-disjoint paths. We formulate an integer linear program and develop heuristics for the problem, and use them to solve various complex and realistic network instances. Numerical data show that, compared to conventional service-provisioning schemes such as multipath routing from a single data center or dedicated-path protection, our proposed scheme efficiently utilizes network resources, improves reliability, and reduces latency; hence, it is suitable for the above-mentioned services.

One-bit quantization, which relies on comparing the signals of interest with given threshold levels, has attracted considerable attention in signal processing for communications and sensing. A useful tool for covariance recovery in such settings is the arcsine law, that estimates the normalized covariance matrix of zero-mean stationary input signals. This relation, however, only considers a zero sampling threshold, which can cause a remarkable information loss. In this paper, the idea of the arcsine law is extended to the case where one-bit analog-to-digital converters (ADCs) apply time-varying thresholds. Specifically, three distinct approaches are proposed, investigated, and compared, to recover the autocorrelation sequence of the stationary signals of interest. Additionally, we will study a modification of the Bussgang law, a famous relation facilitating the recovery of the cross-correlation between the one-bit sampled data and the zero-mean stationary input signal. Similar to the case of the arcsine law, the Bussgang law only considers a zero sampling threshold. This relation is also extended to accommodate the more general case of time-varying thresholds for the stationary input signals.

The recovery of the input signal covariance values from its one-bit sampled counterpart has been deemed a challenging task in the literature. To deal with its difficulties, some assumptions are typically made to find a relation between the input covariance matrix and the autocorrelation values of the one-bit sampled data. This includes the arcsine law and the modified arcsine law that were discussed in Part I of this work [2]. We showed that by facilitating the deployment of time-varying thresholds, the modified arcsine law has a promising performance in covariance recovery. However, the modified arcsine law also assumes input signals are stationary, which is typically a simplifying assumption for real-world applications. In fact, in many signal processing applications, the input signals are readily known to be non-stationary with a non-Toeplitz covariance matrix. In this paper, we propose an approach to extending the arcsine law to the case where one-bit ADCs apply time-varying thresholds while dealing with input signals that originate from a non-stationary process. In particular, the recovery methods are shown to accurately recover the time-varying variance and autocorrelation values. Furthermore, we extend the formulation of the Bussgang law to the case where non-stationary input signals are considered.

A document by Marko Orescanin. Click on the document to view its contents.

A document by Mohammad Ebrahimi . Click on the document to view its contents.

The following report was undertaken to investigate the detailed study for Organic Solar cells with the material review, benefits, and challenges to fulfill the energy demand and give flexibility for the future generation.

Oscillatory Neural Network (ONN) is an emerging neuromorphic architecture composed of oscillators that implement neurons and coupled by synapses. ONNs exhibit rich dynamics and associative properties, which can be used to solve problems in the analog domain according to the paradigm let physics compute. For example, compact oscillators made of VO2 material are good candidates for building low-power ONN architectures dedicated to AI applications at the edge like pattern recognition. However, little is known about the ONN scalability and its performances when implemented in hardware. Before deploying ONN, it is necessary to assess its computation time, energy consumption, performance and accuracy for a given application. Here, we consider a VO2-oscillator as an ONN building block and we perform circuit-level simulations to evaluate the ONN performances at the architecture level. Notably, we investigate how ONN computation time, energy and memory capacity scale with the number of oscillators. We show that ONN energy grows linearly when scaling up the network, making it suitable for large-scale integration at the edge. Furthermore, we investigate the design knobs for minimizing the ONN energy. Assisted by TCAD simulations, we report on scaling of VO2 devices in crossbar geometry to decrease the oscillator voltage and energy. We benchmark ONN versus state-of-the-art architectures and observe that the ONN paradigm is a competitive energy efficient solution for scaled VO2 devices. Finally, we present how ONN can efficiently detect edges in images captured on low-power edge devices.

The stored energies of time-harmonic field in a medium are traditionally analyzed by using narrow-band (NB) approximation. The analysis, however, suffers from two problems. Firstly, it involves a specific constitutive relation and is thus only applicable to a specified medium. Secondly, it is incompatible with the time-harmonic field theory for the stored energy expressions derived from it include time-dependent complex envelops rather than phasors. This article aims to solve the problems with the traditional NB analysis. The expressions of the stored energies in an arbitrary medium have been derived by introducing a new NB technique, which is based on the understanding that a pure sinusoidal field may be viewed as the limit of a NB field as its bandwidth shrinks to zero. To facilitate the derivation, an explicit form of the complex envelope is first deduced by expanding the Fourier frequency spectrum of the NB field around the mid-band frequency. The inflow power from the external source can be divided into the sum of the rate of dissipated energy and the rate of increase of stored energy. After introducing the time-domain fields derived from the new NB approach, the rates of dissipated energies and the stored energies can be determined.

Facial expression recognition, deep convolutional neural network, basic expressions, compound expressions.

In this work, we report guided-mode resonance enhanced room-temperature mid-wave infrared photodetectors with significant spectral-selective enhancement in responsivity and detectivity, all in the absorber thickness of only 250 nm.

This study aimed to evaluate the visual-spatiotemporal co-registration of the real and virtual objects’ movement dynamics in a low-cost physics-based virtual reality (VR) system that provides real cutaneous and kinesthetic haptic feedback of the objects instead of computer-generated haptic feedback. Twelve healthy participants performed three human-robot collaborative (HRC) sequential pick-and-place lifting tasks while both motion capture and VR systems respectively traced the movement kinematics of the real and virtual objects. We used an iterative closest point algorithm to transform the 3D spatial point clouds of the VR system into the motion capture system. We employed a novel algorithm and principal component analysis to respectively calculate visual and spatiotemporal co-registration precisions between virtual and real objects. Results showed a high correlation (r > 0.96) between real and virtual objects’ movement dynamics and linear and angular co-registration errors of less than 5 cm and 8°, respectively. The trend also revealed a low temporal registration error of <12 ms and was only found along the vertical axis. The visual registration data indicated that the real cutaneous and kinesthetic haptics provided by the physical objects in the virtual environment enhanced proprioception and visuomotor functions of the users.

Negotiation is a complex social interaction that encapsulates emotional encounters in human decision-making. Virtual agents that can negotiate with humans are useful in pedagogy and conversational AI. To advance the development of such agents, we explore the prediction of two important subjective goals in a negotiation - outcome satisfaction and partner perception. Specifically, we analyze the extent to which emotion attributes extracted from the negotiation help in the prediction, above and beyond the individual difference variables. We focus on a recent dataset in chat-based negotiations, grounded in a realistic camping scenario. We draw extensive qualitative and quantitative comparisons between three types of emotion variables - emoticons, as well as lexical and contextual}variables, by leveraging affective lexicons and a state-of-the-art deep learning architecture. To further validate the findings, we analyze the prediction of these subjective negotiation goals after controlling for the objective performance of the participants. We also study the temporal effects, understanding the contribution of emotion expressed in the initial and latter parts of the conversation. Finally, we discuss our insights, which will be helpful for designing adaptive negotiation agents that interact through realistic communication interfaces.

A document by Adrian David Cheok . Click on the document to view its contents.

A Computing System for Measuring the Relationship of Smell and Pain

Cooperative multi-agent manipulation systems allow to extend on the manipulative limitations of individual agents, increasing the complexity of the manipulation tasks the ensemble can handle. Controlling such a system requires meticulous planning of subsequent subtasks, queried to the individual agents, in order to execute the master task successfully. Real-time planning is essential to ensure the task can still be achieved when subtasks execution suffers from uncertainty or when the master task changes intermittently requiring real-time reconfiguration of the plan. In this work we develop a supervisory control architecture tailored to the cooperation of two robotic manipulators equipped with standard pick-and-place facilities in the plane. We consider a toy task description where we control the planar position and orientation of an object. A time-invariant policy function is trained using deep reinforcement learning, which can determine a finite sequence pick-and-place maneuvers to manipulate the object into its desired configuration. A comparison is made between two strategies, with the distinction made based on different treatments of the final step. The more information is given to the policy the easier it trains. In return, it becomes less adaptable and loses some of its generalizability.

A document by Ryman Hashem . Click on the document to view its contents.

Implementation of a two Degrees of Freedom Pointing System model to enable ship-satellite communication with the aim of creating self driving cargo ships. The Pointing System is similar to a telescope in terms of Degrees of Freedom (Azimuth and Elevation) and the model kinematics definition was possible using MSC ADAMS software. The model was exported to a Simulink block in order to generate the control block diagram and than an ADAMS-Simulink Co-simulation. The acquisition and orbital propagation errors for the satellite and the acquisition errors of ship motions have been taken into consideration in the control system generation. The main goal is to evaluate the pointing motion to verify the satellite’s locking with the adopted optical communication technologies. In order to reduce pointing error, scanning strategies were considered and the scanning amplitude and frequency values used for the choice of piezoelectric actuators for the optical component.