arXiv subjects

Tao He

Publications and source records attributed to Tao He.

At least 109 records · Page 6Linked to original sources

arXiv2021

Semantic Compositional Learning for Low-shot Scene Graph Generation

Scene graphs provide valuable information to many downstream tasks. Many scene graph generation (SGG) models solely use the limited annotated relation triples for training, leading to their underperformance on low-shot (few and zero) scenarios, especially on the rare predicates. To address this problem, we propose a novel semantic compositional learning strategy that makes it possible to construct additional, realistic relation triples with objects from different images. Specifically, our strategy decomposes a relation triple by identifying and removing the unessential component and composes a new relation triple by fusing with a semantically or visually similar object from a visual components dictionary, whilst ensuring the realisticity of the newly composed triple. Notably, our strategy is generic and can be combined with existing SGG models to significantly improve their performance. We performed a comprehensive evaluation on the benchmark dataset Visual Genome. For three recent SGG models, adding our strategy improves their performance by close to 50\%, and all of them substantially exceed the current state-of-the-art.

cs.CV↗

arXiv2021

The Multi-Modal Video Reasoning and Analyzing Competition

In this paper, we introduce the Multi-Modal Video Reasoning and Analyzing Competition (MMVRAC) workshop in conjunction with ICCV 2021. This competition is composed of four different tracks, namely, video question answering, skeleton-based action recognition, fisheye video-based action recognition, and person re-identification, which are based on two datasets: SUTD-TrafficQA and UAV-Human. We summarize the top-performing methods submitted by the participants in this competition and show their results achieved in the competition.

cs.CV↗

arXiv2021

A Computer-Aided Diagnosis System for Breast Pathology: A Deep Learning Approach with Model Interpretability from Pathological Perspective

Objective: We develop a computer-aided diagnosis (CAD) system using deep learning approaches for lesion detection and classification on whole-slide images (WSIs) with breast cancer. The deep features being distinguishing in classification from the convolutional neural networks (CNN) are demonstrated in this study to provide comprehensive interpretability for the proposed CAD system using pathological knowledge. Methods: In the experiment, a total of 186 slides of WSIs were collected and classified into three categories: Non-Carcinoma, Ductal Carcinoma in Situ (DCIS), and Invasive Ductal Carcinoma (IDC). Instead of conducting pixel-wise classification into three classes directly, we designed a hierarchical framework with the multi-view scheme that performs lesion detection for region proposal at higher magnification first and then conducts lesion classification at lower magnification for each detected lesion. Results: The slide-level accuracy rate for three-category classification reaches 90.8% (99/109) through 5-fold cross-validation and achieves 94.8% (73/77) on the testing set. The experimental results show that the morphological characteristics and co-occurrence properties learned by the deep learning models for lesion classification are accordant with the clinical rules in diagnosis. Conclusion: The pathological interpretability of the deep features not only enhances the reliability of the proposed CAD system to gain acceptance from medical specialists, but also facilitates the development of deep learning frameworks for various tasks in pathology. Significance: This paper presents a CAD system for pathological image analysis, which fills the clinical requirements and can be accepted by medical specialists with providing its interpretability from the pathological perspective.

eess.IV↗

arXiv2021

Design of a 200-MHz Continuous-Wave Radio Frequency Quadrupole Accelerator for Boron Neutron Capture Therapy

A high-intensity continuous wave (CW) radio frequency quadrupole (RFQ) accelerator is designed for boron neutron capture therapy (BNCT). The transmission efficiency of a 20-mA proton beam accelerated from 30 keV to 2.5 MeV can reach 98.7% at an operating frequency of 200 MHz. The beam dynamics has good tolerance to errors. By comparing the high-frequency parameters of quadrilateral and octagonal RFQ cross-sections, the quadrilateral structure of the four-vane cavity is selected owing to its multiple advantages, such as a smaller cross section at the same frequency and easy processing. In addition, tuners and undercuts are designed to tune the frequency of the cavity and achieve a flat electric field distribution along the cavity. In this paper, the beam dynamics simulation and electromagnetic design are presented in detail.

physics.acc-ph↗

arXiv2020

R3Det: Refined Single-Stage Detector with Feature Refinement for Rotating Object

Rotation detection is a challenging task due to the difficulties of locating the multi-angle objects and separating them effectively from the background. Though considerable progress has been made, for practical settings, there still exist challenges for rotating objects with large aspect ratio, dense distribution and category extremely imbalance. In this paper, we propose an end-to-end refined single-stage rotation detector for fast and accurate object detection by using a progressive regression approach from coarse to fine granularity. Considering the shortcoming of feature misalignment in existing refined single-stage detector, we design a feature refinement module to improve detection performance by getting more accurate features. The key idea of feature refinement module is to re-encode the position information of the current refined bounding box to the corresponding feature points through pixel-wise feature interpolation to realize feature reconstruction and alignment. For more accurate rotation estimation, an approximate SkewIoU loss is proposed to solve the problem that the calculation of SkewIoU is not derivable. Experiments on three popular remote sensing public datasets DOTA, HRSC2016, UCAS-AOD as well as one scene text dataset ICDAR2015 show the effectiveness of our approach. Tensorflow and Pytorch version codes are available at https://github.com/Thinklab-SJTU/R3Det_Tensorflow and https://github.com/SJTU-Thinklab-Det/r3det-on-mmdetection, and R3Det is also integrated in our open source rotation detection benchmark: https://github.com/yangxue0827/RotationDetection.

cs.CV↗

arXiv2020

Learning from the Scene and Borrowing from the Rich: Tackling the Long Tail in Scene Graph Generation

Despite the huge progress in scene graph generation in recent years, its long-tail distribution in object relationships remains a challenging and pestering issue. Existing methods largely rely on either external knowledge or statistical bias information to alleviate this problem. In this paper, we tackle this issue from another two aspects: (1) scene-object interaction aiming at learning specific knowledge from a scene via an additive attention mechanism; and (2) long-tail knowledge transfer which tries to transfer the rich knowledge learned from the head into the tail. Extensive experiments on the benchmark dataset Visual Genome on three tasks demonstrate that our method outperforms current state-of-the-art competitors.

cs.CV↗

arXiv2019

One Network for Multi-Domains: Domain Adaptive Hashing with Intersectant Generative Adversarial Network

With the recent explosive increase of digital data, image recognition and retrieval become a critical practical application. Hashing is an effective solution to this problem, due to its low storage requirement and high query speed. However, most of past works focus on hashing in a single (source) domain. Thus, the learned hash function may not adapt well in a new (target) domain that has a large distributional difference with the source domain. In this paper, we explore an end-to-end domain adaptive learning framework that simultaneously and precisely generates discriminative hash codes and classifies target domain images. Our method encodes two domains images into a semantic common space, followed by two independent generative adversarial networks arming at crosswise reconstructing two domains' images, reducing domain disparity and improving alignment in the shared space. We evaluate our framework on {four} public benchmark datasets, all of which show that our method is superior to the other state-of-the-art methods on the tasks of object recognition and image retrieval.

cs.CV↗

arXiv2019

Understanding the Mechanism of Deep Learning Framework for Lesion Detection in Pathological Images with Breast Cancer

The computer-aided detection (CADe) systems are developed to assist pathologists in slide assessment, increasing diagnosis efficiency and reducing missing inspections. Many studies have shown such a CADe system with deep learning approaches outperforms the one using conventional methods that rely on hand-crafted features based on field-knowledge. However, most developers who adopted deep learning models directly focused on the efficacy of outcomes, without providing comprehensive explanations on why their proposed frameworks can work effectively. In this study, we designed four experiments to verify the consecutive concepts, showing that the deep features learned from pathological patches are interpretable by domain knowledge of pathology and enlightening for clinical diagnosis in the task of lesion detection. The experimental results show the activation features work as morphological descriptors for specific cells or tissues, which agree with the clinical rules in classification. That is, the deep learning framework not only detects the distribution of tumor cells but also recognizes lymphocytes, collagen fibers, and some other non-cell structural tissues. Most of the characteristics learned by the deep learning models have summarized the detection rules that can be recognized by the experienced pathologists, whereas there are still some features may not be intuitive to domain experts but discriminative in classification for machines. Those features are worthy to be further studied in order to find out the reasonable correlations to pathological knowledge, from which pathological experts may draw inspirations for exploring new characteristics in diagnosis.

cs.CV↗

arXiv2018

Photocorrosion-limited maximum efficiency of solar photoelectrochemical water splitting

Photoelectrochemical (PEC) water splitting to generate hydrogen is one of the most studied methods for converting solar energy into clean fuel because of its simplicity and potentially low cost. Despite over 40 years of intensive research, PEC water splitting remains in its early stages with stable efficiencies far less than 10%, a benchmark for commercial applications. Here, we revealed that the desired photocorrosion stability sets a limit of 2.48 eV (relative to the normal hydrogen electrode (NHE)) for the highest possible potential of the valence band (VB) edge of a photocorrosion-resistant semiconducting photocatalyst. We further demonstrated that such limitation has a deep root in underlying physics after deducing the relation between energy position of the valence band edge and free-energy for a semiconductor. The disparity between the stability-limited VB potential at 2.48 V and the oxygen evolution reaction (OER) potential at 1.23 V vs NHE reduces the maximum STH conversion efficiency to approximately 8% for long-term stable single-bandgap PEC water splitting cells. Based on this understanding, we suggest that the most promising strategy to overcome this 8% efficiency limit is to decouple the requirements of efficient light harvesting and chemical stability by protecting the active semiconductor photocatalyst surface with a photocorrosion-resistant oxide coating layer.

cond-mat.mtrl-sci↗

arXiv2017

Deep Discrete Hashing with Self-supervised Pairwise Labels

Hashing methods have been widely used for applications of large-scale image retrieval and classification. Non-deep hashing methods using handcrafted features have been significantly outperformed by deep hashing methods due to their better feature representation and end-to-end learning framework. However, the most striking successes in deep hashing have mostly involved discriminative models, which require labels. In this paper, we propose a novel unsupervised deep hashing method, named Deep Discrete Hashing (DDH), for large-scale image retrieval and classification. In the proposed framework, we address two main problems: 1) how to directly learn discrete binary codes? 2) how to equip the binary representation with the ability of accurate image retrieval and classification in an unsupervised way? We resolve these problems by introducing an intermediate variable and a loss function steering the learning process, which is based on the neighborhood structure in the original space. Experimental results on standard datasets (CIFAR-10, NUS-WIDE, and Oxford-17) demonstrate that our DDH significantly outperforms existing hashing methods by large margin in terms of~mAP for image retrieval and object recognition. Code is available at \url{https://github.com/htconquer/ddh}.

cs.CV↗

arXiv2017

A Fast Proximal Gradient Algorithm for Decentralized Composite Optimization over Directed Networks

This paper proposes a fast decentralized algorithm for solving a consensus optimization problem defined in a directed networked multi-agent system, where the local objective functions have the smooth+nonsmooth composite form, and are possibly nonconvex. Examples of such problems include decentralized compressed sensing and constrained quadratic programming problems, as well as many decentralized regularization problems. We extend the existing algorithms PG-EXTRA and ExtraPush to a new algorithm PG-ExtraPush for composite consensus optimization over a directed network. This algorithm takes advantage of the proximity operator like in PG-EXTRA to deal with the nonsmooth term, and employs the push-sum protocol like in ExtraPush to tackle the bias introduced by the directed network. With a proper step size, we show that PG-ExtraPush converges to an optimal solution at a linear rate under some regular assumptions. We conduct a series of numerical experiments to show the effectiveness of the proposed algorithm. Specifically, with a proper step size, PG-ExtraPush performs linear rates in most of cases, even in some nonconvex cases, and is significantly faster than Subgradient-Push, even if the latter uses a hand-optimized step size. The established theoretical results are also verified by the numerical results.

cs.DC↗

arXiv2017

Deep Region Hashing for Efficient Large-scale Instance Search from Images

Instance Search (INS) is a fundamental problem for many applications, while it is more challenging comparing to traditional image search since the relevancy is defined at the instance level. Existing works have demonstrated the success of many complex ensemble systems that are typically conducted by firstly generating object proposals, and then extracting handcrafted and/or CNN features of each proposal for matching. However, object bounding box proposals and feature extraction are often conducted in two separated steps, thus the effectiveness of these methods collapses. Also, due to the large amount of generated proposals, matching speed becomes the bottleneck that limits its application to large-scale datasets. To tackle these issues, in this paper we propose an effective and efficient Deep Region Hashing (DRH) approach for large-scale INS using an image patch as the query. Specifically, DRH is an end-to-end deep neural network which consists of object proposal, feature extraction, and hash code generation. DRH shares full-image convolutional feature map with the region proposal network, thus enabling nearly cost-free region proposals. Also, each high-dimensional, real-valued region features are mapped onto a low-dimensional, compact binary codes for the efficient object region level matching on large-scale dataset. Experimental results on four datasets show that our DRH can achieve even better performance than the state-of-the-arts in terms of MAP, while the efficiency is improved by nearly 100 times.

cs.CV↗

arXiv2009

Resistive switching in nanogap systems on SiO2 substrates

Voltage-controlled resistive switching is demonstrated in various gap systems on SiO2 substrate. The nanosized gaps are made by different means using different materials including metal, semiconductor, and metallic nonmetal. The switching site is further reduced by using multi-walled carbon nanotubes and single-walled carbon nanotubes. The switching in all the gap systems shares the same characteristics. This independence of switching on the material compositions of the electrodes, accompanied by observable damage to the SiO2 substrate at the gap region, bespeaks the intrinsic switching from post-breakdown SiO2. It calls for caution when studying resistive switching in nanosystems on oxide substrates, since oxide breakdown extrinsic to the nanosystem can mimic resistive switching. Meanwhile, the high ON/OFF ratio (10E5), fast switching time (2 us, test limit), durable cycles demonstrated show promising memory properties. The intermediate states observed reveal the filamentary conduction nature.

cond-mat.mes-hall↗

arXiv2009

Orientation-dependent stability and quantum-confinement effects of silicon carbide nanowires

The energetic stability and electronic properties of hydrogenated silicon carbide nanowires (SiCNWs) with zinc blende (3C) and wurtzite (2H) structures are investigated using first-principles calculations within density functional theory and generalized gradient approximation. The [111]-orientated 3C-SiCNWs are energetically more stable than other kinds of NWs with similar size. All the NWs have direct band gaps except the 3C-SiCNWs orientating along [112] direction. The band gaps of these NWs decrease with the increase of wire size, due to the quantum-confinement effects. The direct-band-gap features can be kept for the 3C-SiCNWs orientating along [111] direction with diameters up to 2.8 nm. The superior stability and electronic structures of the [111]-orientated 3C-SiCNWs are in good agreement with the experimental results.

cond-mat.mtrl-sci↗

arXiv2008

Controlling transistor threshold voltages using molecular dipoles

We develop a theoretical model for how organic molecules can control the electronic and transport properties of an underlying transistor channel to whose surface they are chemically bonded. The influence arises from a combination of long-ranged dipolar electrostatics due to the molecular head-groups, as well as short-ranged charge transfer and interfacial dipole driven by equilibrium band-alignment between the molecular backbone and the reconstructed semiconductor surface atoms.

cond-mat.mes-hall↗

arXiv2004

Coupling between electronic and structural degrees of freedom in the triangular lattice conductor NaxCoO2

The determination by powder neutron diffraction of the ambient temperature crystal structures of compounds in the NaxCoO2 family, for 0.3 < x <= 1.0, is reported. The structures consist of triangular CoO2 layers with Na ions distributed in intervening charge reservoir layers. The shapes of the CoO6 octahedra that make up the CoO2 layers are found to be critically dependent on the electron count and on the distribution of the Na ions in the intervening layers, where two types of Na sites are available. Correlation of the shapes of cobalt-oxygen octahedra, the Na ion positions, and the electronic phase diagram in NaxCoO2 is made, showing how structural and electronic degrees of freedom can be coupled in electrically conducting triangular lattice systems.

cond-mat.str-el↗

arXiv2004

An experimental and computational investigation of structure and magnetism in pyrite Co$_{1-x}$Fe$_x$S$_2$: Chemical bonding and half-metallicity

Bulk samples of the pyrite chalcogenide solid solutions Co$_{1-x}$Fe$_x$S$_2$ 0 <= x <= 0.5, have been prepared and their crystal structures and magnetic properties studied by X-ray diffraction and SQUID magnetization measurements. Across the solution series, the distance between sulfur atoms in the persulfide (S$_2^{2-}$) unit remains nearly constant. First principles electronic structure calculations using experimental crystal structures as inputs point to the importance of this constant S-S distance, in helping antibonding S-S levels pin the Fermi energy. In contrast hypothetical rock-salt CoS is not a good half metal, despite being nearly isostructural and isoelectronic. We use our understanding of the Co$_{1-x}$Fe$_x$S$_2$ system to make some prescriptions for new ferromagnetic half-metals.

cond-mat.mtrl-sci↗

arXiv2004

Charge-ordering, commensurability and metallicity in the phase diagram of layered Na(x)CoO(2)

The phase diagram of non-hydrated Na(x)CoO(2) has been determined by changing the Na content x using a series of chemical reactions. As x increases from 0.3, the ground state goes from a paramagnetic metal to a charge-ordered insulator (at x=1/2) to a `Curie-Weiss metal' (around 0.70), and finally to a weak-moment magnetically ordered state (x>0.75). The unusual properties of the state at 1/2 (including particle-hole symmetry at low T and enhanced thermal conductivity) are described. The strong coupling between the Na ions and the holes is emphasized.

cond-mat.supr-con↗