arXiv subjects

Di Huang

Publications and source records attributed to Di Huang.

At least 199 records · Page 11Linked to original sources

arXiv2022

Racial Disparities in Pulse Oximetry Cannot Be Fixed With Race-Based Correction

Studies have shown pulse oximeter measurements of blood oxygenation have statistical bias that is a function of race, which results in higher rates of occult hypoxemia, i.e., missed detection of dangerously low oxygenation, in patients of color. This paper further characterizes the statistical distribution of pulse ox measurements, showing they also have a higher variance for patients racialized as Black, compared to those racialized as white. We show that no single race-based correction factor will provide equal performance in the detection of hypoxemia. The results have implications for racially equitable pulse oximetry.

eess.SP↗

arXiv2022

Phonon-assisted inter-valley scattering determines ultrafast exciton dynamics in MoSe$_2$ bilayers

While valleys (energy extrema) are present in all band structures of solids, their preeminent role in determining exciton resonances and dynamics in atomically thin transition metal dichalcogenides (TMDC) is unique. Using two-dimensional coherent electronic spectroscopy, we find that exciton decoherence occurs on a much faster time scale in MoSe$_2$ bilayers than that in the monolayers. We further identify two population relaxation channels in the bilayer, a coherent and an incoherent one. Our microscopic model reveals that phonon-emission processes facilitate scattering events from the $K$ valley to other lower energy $Γ$ and $Λ$ valleys in the bilayer. Our combined experimental and theoretical studies unequivocally establish different microscopic mechanisms that determine exciton quantum dynamics in TMDC monolayers and bilayers. Understanding exciton quantum dynamics provides critical guidance to manipulation of spin/valley degrees of freedom in TMDC bilayers.

cond-mat.mes-hall↗

arXiv2022

STS: Surround-view Temporal Stereo for Multi-view 3D Detection

Learning accurate depth is essential to multi-view 3D object detection. Recent approaches mainly learn depth from monocular images, which confront inherent difficulties due to the ill-posed nature of monocular depth learning. Instead of using a sole monocular depth method, in this work, we propose a novel Surround-view Temporal Stereo (STS) technique that leverages the geometry correspondence between frames across time to facilitate accurate depth learning. Specifically, we regard the field of views from all cameras around the ego vehicle as a unified view, namely surroundview, and conduct temporal stereo matching on it. The resulting geometrical correspondence between different frames from STS is utilized and combined with the monocular depth to yield final depth prediction. Comprehensive experiments on nuScenes show that STS greatly boosts 3D detection ability, notably for medium and long distance objects. On BEVDepth with ResNet-50 backbone, STS improves mAP and NDS by 2.6% and 1.4%, respectively. Consistent improvements are observed when using a larger backbone and a larger image resolution, demonstrating its effectiveness

cs.CV↗

arXiv2022

Motion Sensitive Contrastive Learning for Self-supervised Video Representation

Contrastive learning has shown great potential in video representation learning. However, existing approaches fail to sufficiently exploit short-term motion dynamics, which are crucial to various down-stream video understanding tasks. In this paper, we propose Motion Sensitive Contrastive Learning (MSCL) that injects the motion information captured by optical flows into RGB frames to strengthen feature learning. To achieve this, in addition to clip-level global contrastive learning, we develop Local Motion Contrastive Learning (LMCL) with frame-level contrastive objectives across the two modalities. Moreover, we introduce Flow Rotation Augmentation (FRA) to generate extra motion-shuffled negative samples and Motion Differential Sampling (MDS) to accurately screen training samples. Extensive experiments on standard benchmarks validate the effectiveness of the proposed method. With the commonly-used 3D ResNet-18 as the backbone, we achieve the top-1 accuracies of 91.5\% on UCF101 and 50.3\% on Something-Something v2 for video classification, and a 65.6\% Top-1 Recall on UCF101 for video retrieval, notably improving the state-of-the-art.

cs.CV↗

arXiv2022

Neutron-induced nuclear recoil background in the PandaX-4T experiment

Neutron-induced nuclear recoil background is critical to the dark matter searches in the PandaX-4T liquid xenon experiment. This paper studies the feature of neutron background in liquid xenon and evaluates their contribution in the single scattering nuclear recoil events through three methods. The first method is fully Monte Carlo simulation based. The last two are data-driven methods that also use the multiple scattering signals and high energy signals in the data, respectively. In the PandaX-4T commissioning data with an exposure of 0.63 tonne-year, all these methods give a consistent result that there are $1.15\pm0.57$ neutron-induced background in dark matter signal region within an approximated nuclear recoil energy window between 5 and 100 keV.

hep-ex↗

arXiv2022

Video Anomaly Detection by Solving Decoupled Spatio-Temporal Jigsaw Puzzles

Video Anomaly Detection (VAD) is an important topic in computer vision. Motivated by the recent advances in self-supervised learning, this paper addresses VAD by solving an intuitive yet challenging pretext task, i.e., spatio-temporal jigsaw puzzles, which is cast as a multi-label fine-grained classification problem. Our method exhibits several advantages over existing works: 1) the spatio-temporal jigsaw puzzles are decoupled in terms of spatial and temporal dimensions, responsible for capturing highly discriminative appearance and motion features, respectively; 2) full permutations are used to provide abundant jigsaw puzzles covering various difficulty levels, allowing the network to distinguish subtle spatio-temporal differences between normal and abnormal events; and 3) the pretext task is tackled in an end-to-end manner without relying on any pre-trained models. Our method outperforms state-of-the-art counterparts on three public benchmarks. Especially on ShanghaiTech Campus, the result is superior to reconstruction and prediction-based methods by a large margin.

cs.CV↗

arXiv2022

Study of background from accidental coincidence signals in the PandaX-II experiment

The PandaX-II experiment employed a 580kg liquid xenon detector to search for the interactions between dark matter particles and the target xenon atoms. The accidental coincidences of isolated signals result in a dangerous background which mimic the signature of the dark matter. We performed a detailed study on the accidental coincidence background in PandaX-II, including the possible origin of the isolated signals, the background level and corresponding background suppression method. With a boosted-decision-tree algorithm, the accidental coincidence background is reduced by 70% in the dark matter signal region, thus the sensitivity of dark matter search at PandaX-II is improved.

hep-ex↗

arXiv2022

Representation Learning for Compressed Video Action Recognition via Attentive Cross-modal Interaction with Motion Enhancement

Compressed video action recognition has recently drawn growing attention, since it remarkably reduces the storage and computational cost via replacing raw videos by sparsely sampled RGB frames and compressed motion cues (e.g., motion vectors and residuals). However, this task severely suffers from the coarse and noisy dynamics and the insufficient fusion of the heterogeneous RGB and motion modalities. To address the two issues above, this paper proposes a novel framework, namely Attentive Cross-modal Interaction Network with Motion Enhancement (MEACI-Net). It follows the two-stream architecture, i.e. one for the RGB modality and the other for the motion modality. Particularly, the motion stream employs a multi-scale block embedded with a denoising module to enhance representation learning. The interaction between the two streams is then strengthened by introducing the Selective Motion Complement (SMC) and Cross-Modality Augment (CMA) modules, where SMC complements the RGB modality with spatio-temporally attentive local motion features and CMA further combines the two modalities with selective feature augmentation. Extensive experiments on the UCF-101, HMDB-51 and Kinetics-400 benchmarks demonstrate the effectiveness and efficiency of MEACI-Net.

cs.CV↗

arXiv2022

A search for two-component Majorana dark matter in a simplified model using the full exposure data of PandaX-II experiment

In the two-component Majorana dark matter model, one dark matter particle can scatter off the target nuclei, and turn into a slightly heavier component. In the framework of a simplified model with a vector boson mediator, both the tree-level and loop-level processes contribute to the signal in direct detection experiment. In this paper, we report the search results for such dark matter from PandaX-II experiment, using total data of the full 100.7 tonne$\cdot$day exposure. No significant excess is observed, so strong constraints on the combined parameter space of mediator mass and dark matter mass are derived. With the complementary search results from collider experiments, a large range of parameter space can be excluded.

hep-ex↗

arXiv2022

Neural Program Synthesis with Query

Aiming to find a program satisfying the user intent given input-output examples, program synthesis has attracted increasing interest in the area of machine learning. Despite the promising performance of existing methods, most of their success comes from the privileged information of well-designed input-output examples. However, providing such input-output examples is unrealistic because it requires the users to have the ability to describe the underlying program with a few input-output examples under the training distribution. In this work, we propose a query-based framework that trains a query neural network to generate informative input-output examples automatically and interactively from a large query space. The quality of the query depends on the amount of the mutual information between the query and the corresponding program, which can guide the optimization of the query framework. To estimate the mutual information more accurately, we introduce the functional space (F-space) which models the relevance between the input-output examples and the programs in a differentiable way. We evaluate the effectiveness and generalization of the proposed query-based framework on the Karel task and the list processing task. Experimental results show that the query-based framework can generate informative input-output examples which achieve and even outperform well-designed input-output examples.

cs.LG↗

arXiv2022

Metasurface electron optics in graphene

For electron optics in graphene, the propagation effect has so far been the only physical mechanism available. The resulting electron-optics-based components are large in size and operate at low temperatures to avoid violating the ballistic transport limits. In this paper, Dirac fermion metasurfaces, electronic counterparts of optical metasurfaces, are introduced for graphene electronics. By a metasurface, formally a linear array of gate-bias-controlled circular quantum dots, the wavefront of electron beams can be shaped within a one-quantum-dot-diameter distance, far below the ballistic limits at room temperature. This provides opportunities to create electron-optics-based devices that operate under ambient conditions. Moreover, unlike optical metasurfaces, Dirac fermion metasurfaces have near-perfect operating efficiencies and their high tunability allows for free and fast switching among functionalities. The concept of metasurface electron optics might open up a promising avenue for improving the performance of quantum devices in Dirac fermion materials.

cond-mat.mes-hall↗

arXiv2022

ImFace: A Nonlinear 3D Morphable Face Model with Implicit Neural Representations

Precise representations of 3D faces are beneficial to various computer vision and graphics applications. Due to the data discretization and model linearity, however, it remains challenging to capture accurate identity and expression clues in current studies. This paper presents a novel 3D morphable face model, namely ImFace, to learn a nonlinear and continuous space with implicit neural representations. It builds two explicitly disentangled deformation fields to model complex shapes associated with identities and expressions, respectively, and designs an improved learning strategy to extend embeddings of expressions to allow more diverse changes. We further introduce a Neural Blend-Field to learn sophisticated details by adaptively blending a series of local fields. In addition to ImFace, an effective preprocessing pipeline is proposed to address the issue of watertight input requirement in implicit representations, enabling them to work with common facial surfaces for the first time. Extensive experiments are performed to demonstrate the superiority of ImFace.

cs.CV↗

arXiv2022

Low Radioactive Material Screening and Background Control for the PandaX-4T Experiment

PandaX-4T is a ton-scale dark matter direct detection experiment using a dual-phase TPC technique at the China Jinping Underground Laboratory. Various ultra-low background technologies have been developed and applied to material screening for PandaX-4T, including HPGe gamma spectroscopy, ICP-MS, NAA, radon emanation measurement system, krypton assay station, and alpha detection system. Low background materials were selected to assemble the detector. Surface treatment procedures were investigated to further suppress radioactive background. Combining measured results and Monte Carlo simulation, the total material background rates of PandaX-4T in the energy region of 1-25 keV$\rm{}_{ee}$ are estimated to be (9.9 $\pm$ 1.9) $\times \ 10^{-3}$ mDRU for electron recoil and (2.8 $\pm$ 0.6) $\times \ 10^{-4}$ mDRU for nuclear recoil. In addition, $^{nat}$Kr in the detector is estimated to be <8 ppt.

physics.ins-det↗

arXiv2022

Target-Relevant Knowledge Preservation for Multi-Source Domain Adaptive Object Detection

Domain adaptive object detection (DAOD) is a promising way to alleviate performance drop of detectors in new scenes. Albeit great effort made in single source domain adaptation, a more generalized task with multiple source domains remains not being well explored, due to knowledge degradation during their combination. To address this issue, we propose a novel approach, namely target-relevant knowledge preservation (TRKP), to unsupervised multi-source DAOD. Specifically, TRKP adopts the teacher-student framework, where the multi-head teacher network is built to extract knowledge from labeled source domains and guide the student network to learn detectors in unlabeled target domain. The teacher network is further equipped with an adversarial multi-source disentanglement (AMSD) module to preserve source domain-specific knowledge and simultaneously perform cross-domain alignment. Besides, a holistic target-relevant mining (HTRM) scheme is developed to re-weight the source images according to the source-target relevance. By this means, the teacher network is enforced to capture target-relevant knowledge, thus benefiting decreasing domain shift when mentoring object detection in the target domain. Extensive experiments are conducted on various widely used benchmarks with new state-of-the-art scores reported, highlighting the effectiveness.

cs.CV↗

arXiv2022

Entropy-based Active Learning for Object Detection with Progressive Diversity Constraint

Active learning is a promising alternative to alleviate the issue of high annotation cost in the computer vision tasks by consciously selecting more informative samples to label. Active learning for object detection is more challenging and existing efforts on it are relatively rare. In this paper, we propose a novel hybrid approach to address this problem, where the instance-level uncertainty and diversity are jointly considered in a bottom-up manner. To balance the computational complexity, the proposed approach is designed as a two-stage procedure. At the first stage, an Entropy-based Non-Maximum Suppression (ENMS) is presented to estimate the uncertainty of every image, which performs NMS according to the entropy in the feature space to remove predictions with redundant information gains. At the second stage, a diverse prototype (DivProto) strategy is explored to ensure the diversity across images by progressively converting it into the intra-class and inter-class diversities of the entropy-based class-specific prototypes. Extensive experiments are conducted on MS COCO and Pascal VOC, and the proposed approach achieves state of the art results and significantly outperforms the other counterparts, highlighting its superiority.

cs.CV↗

arXiv2022

A Search for the Cosmic Ray Boosted Sub-GeV Dark Matter at the PandaX-II Experiment

We report a novel search for the cosmic ray boosted dark matter using the 100~tonne$\cdot$day full data set of the PandaX-II detector located at the China Jinping Underground Laboratory. With the extra energy gained from the cosmic rays, sub-GeV dark matter particles can produce visible recoil signals in the detector. The diurnal modulations in rate and energy spectrum are utilized to further enhance the signal sensitivity. Our result excludes the dark matter-nucleon elastic scattering cross section between 10$^{-31}$cm$^{2}$ and 10$^{-28}$cm$^{2}$ for a dark matter masses from 0.1 MeV/$c^2$ to 0.1 GeV/$c^2$, with a large parameter space previously unexplored by experimental collaborations.

hep-ex↗

arXiv2022

Beyond 3DMM: Learning to Capture High-fidelity 3D Face Shape

3D Morphable Model (3DMM) fitting has widely benefited face analysis due to its strong 3D priori. However, previous reconstructed 3D faces suffer from degraded visual verisimilitude due to the loss of fine-grained geometry, which is attributed to insufficient ground-truth 3D shapes, unreliable training strategies and limited representation power of 3DMM. To alleviate this issue, this paper proposes a complete solution to capture the personalized shape so that the reconstructed shape looks identical to the corresponding person. Specifically, given a 2D image as the input, we virtually render the image in several calibrated views to normalize pose variations while preserving the original image geometry. A many-to-one hourglass network serves as the encode-decoder to fuse multiview features and generate vertex displacements as the fine-grained geometry. Besides, the neural network is trained by directly optimizing the visual effect, where two 3D shapes are compared by measuring the similarity between the multiview images rendered from the shapes. Finally, we propose to generate the ground-truth 3D shapes by registering RGB-D images followed by pose and shape augmentation, providing sufficient data for network training. Experiments on several challenging protocols demonstrate the superior reconstruction accuracy of our proposal on the face shape.

cs.CV↗

arXiv2022

CAT-Det: Contrastively Augmented Transformer for Multi-modal 3D Object Detection

In autonomous driving, LiDAR point-clouds and RGB images are two major data modalities with complementary cues for 3D object detection. However, it is quite difficult to sufficiently use them, due to large inter-modal discrepancies. To address this issue, we propose a novel framework, namely Contrastively Augmented Transformer for multi-modal 3D object Detection (CAT-Det). Specifically, CAT-Det adopts a two-stream structure consisting of a Pointformer (PT) branch, an Imageformer (IT) branch along with a Cross-Modal Transformer (CMT) module. PT, IT and CMT jointly encode intra-modal and inter-modal long-range contexts for representing an object, thus fully exploring multi-modal information for detection. Furthermore, we propose an effective One-way Multi-modal Data Augmentation (OMDA) approach via hierarchical contrastive learning at both the point and object levels, significantly improving the accuracy only by augmenting point-clouds, which is free from complex generation of paired samples of the two modalities. Extensive experiments on the KITTI benchmark show that CAT-Det achieves a new state-of-the-art, highlighting its effectiveness.

cs.CV↗