arXiv subjects

Yi Yang

Publications and source records attributed to Yi Yang.

At least 1,027 records · Page 57Linked to original sources

arXiv2018

Supervision-by-Registration: An Unsupervised Approach to Improve the Precision of Facial Landmark Detectors

In this paper, we present supervision-by-registration, an unsupervised approach to improve the precision of facial landmark detectors on both images and video. Our key observation is that the detections of the same landmark in adjacent frames should be coherent with registration, i.e., optical flow. Interestingly, the coherency of optical flow is a source of supervision that does not require manual labeling, and can be leveraged during detector training. For example, we can enforce in the training loss function that a detected landmark at frame$_{t-1}$ followed by optical flow tracking from frame$_{t-1}$ to frame$_t$ should coincide with the location of the detection at frame$_t$. Essentially, supervision-by-registration augments the training loss function with a registration loss, thus training the detector to have output that is not only close to the annotations in labeled images, but also consistent with registration on large amounts of unlabeled videos. End-to-end training with the registration loss is made possible by a differentiable Lucas-Kanade operation, which computes optical flow registration in the forward pass, and back-propagates gradients that encourage temporal coherency in the detector. The output of our method is a more precise image-based facial landmark detector, which can be applied to single images or video. With supervision-by-registration, we demonstrate (1) improvements in facial landmark detection on both images (300W, ALFW) and video (300VW, Youtube-Celebrities), and (2) significant reduction of jittering in video detections.

cs.CV↗

arXiv2018

Subaru/HiCIAO $HK_{\rm s}$ imaging of LkH$α$ 330 - multi-band detection of the gap and spiral-like structures

We present $H$- and $K_{\rm s}$-bands observations of the LkH$α$ 330 disk with a multi-band detection of the large gap and spiral-like structures. The morphology of the outer disk ($r\sim$$0\farcs3$) at PA=0--45$^\circ$ and PA=180--290$^\circ$ are likely density wave-induced spirals and comparison between our observational results and simulations suggests a planet formation. We have also investigated the azimuthal profiles at the ring and the outer-disk regions as well as radial profiles in the directions of the spiral-like structures and semi-major axis. Azimuthal analysis shows a large variety in wavelength and implies that the disk has non-axisymmetric dust distributions. The radial profiles in the major-axis direction (PA=$271^\circ$) suggest that the outer region ($r\geq0\farcs25$) may be influenced by shadows of the inner region of the disk. The spiral-like directions (PA=10$^\circ$ and 230$^\circ$) show different radial profiles, which suggests that the surfaces of the spiral-like structures are highly flared and/or have different dust properties. Finally, a color-map of the disk shows a lack of an outer eastern region in the $H$-band disk, which may hint the presence of an inner object that casts a directional shadow onto the disk.

astro-ph.EP↗

arXiv2018

Holographic Entanglement Entropy in Boundary Quantum Field Theory

We study the holographic entanglement entropy in a (d+1)-dimensional boundary quantum field theory at both the zero and finite temperature. The phase diagrams for the holographic entanglement entropy at various temperatures are obtained by solving the entangled surfaces in the different homology. We also verify the Araki-Lieb inequality and illustrate the entanglement plateau.

hep-th↗

arXiv2018

Subdivisions of the posteromedial cortex in disorders of consciousness

Evidence suggests that disruptions of the posteromedial cortex (PMC) and posteromedial corticothalamic connectivity contribute to disorders of consciousness (DOCs). While most previous studies treated the PMC as a whole, this structure is functionally heterogeneous. The present study investigated whether particular subdivisions of the PMC are specifically associated with DOCs. Participants were DOC patients, 21 vegetative state/unresponsive wakefulness syndrome (VS/UWS), 12 minimally conscious state (MCS), and 29 healthy controls. Individual PMC and thalamus were divided into distinct subdivisions by their fiber tractograpy to each other and default mode regions, and white matter integrity and brain activity between/within subdivisions were assessed. The thalamus was represented mainly in the dorsal and posterior portions of the PMC, and the white matter tracts connecting these subdivisions to the thalamus had less integrity in VS/UWS patients than in MCS patients and healthy controls, as well as in patients who did not recover after 12 months than in patients who did. The structural substrates were validated by finding impaired functional fluctuations within this PMC subdivision. This study is the first to show that tracts from dorsal and posterior subdivisions of the PMC to the thalamus contribute to DOCs.

q-bio.NC↗

arXiv2018

Deploy Large-Scale Deep Neural Networks in Resource Constrained IoT Devices with Local Quantization Region

Implementing large-scale deep neural networks with high computational complexity on low-cost IoT devices may inevitably be constrained by limited computation resource, making the devices hard to respond in real-time. This disjunction makes the state-of-art deep learning algorithms, i.e. CNN (Convolutional Neural Networks), incompatible with IoT world. We present a low-bit (range from 8-bit to 1-bit) scheme with our local quantization region algorithm. We use models in Caffe model zoo as our example tasks to evaluate the effect of our low precision data representation scheme. With the available of local quantization region, we find implementations on top of those schemes could greatly retain the model accuracy, besides the reduction of computational complexity. For example, our 8-bit scheme has no drops on top-1 and top-5 accuracy with 2x speedup on Intel Edison IoT platform. Implementations based on our 4-bit, 2-bit or 1-bit scheme are also applicable to IoT devices with advances of low computational complexity. For example, the drop on our task is only 0.7% when using 2-bit scheme, a scheme which could largely save transistors. Making low-bit scheme usable here opens a new door for further optimization on commodity IoT controller, i.e. extra speed-up could be achieved by replacing multiply-accumulate operations with the proposed table look-up operations. The whole study offers a new approach to relief the challenge of bring advanced deep learning algorithm to resource constrained low-cost IoT device.

cs.LG↗

arXiv2018

High-contrast Polarimetry Observation of T Tau Circumstellar Environment

We conducted high-contrast polarimetry observations of T Tau in the H-band, using the HiCIAO instrument mounted on the Subaru Telescope, revealing structures as near as 0.$\arcsec$1 from the stars T Tau N and T Tau S. The whole T Tau system is found to be surrounded by nebula-like envelopes, and several outflow-related structures are detected in these envelopes. We analyzed the detailed polarization patterns of the circumstellar structures near each component of this triple young star system and determined constraints on the circumstellar disks and outflow structures. We suggest that the nearly face-on circumstellar disk of T Tau N is no larger than 0.$\arcsec$8, or 117 AU, in the northwest, based on the existence of a hole in this direction, and no larger than 0.$\arcsec$27, or 40 AU, in the south. A new structure "N5" extends to about 0.$\arcsec$42, or 59 AU, on the southwest of the star, believed to be part of the disk. We suggest that T Tau S is surrounded by a highly inclined circumbinary disk with a radius of about 0.$\arcsec$3, or 44 AU, with a position angle of about 30$^\circ$, that is misaligned with the orbit of the T Tau S binary. After analyzing the positions and polarization vector patterns of the outflow-related structures, we suggest that T Tau S should trigger the well-known E-W outflow, and is also likely to be responsible for a southwest precessing outflow "coil" and a possible south outflow.

astro-ph.SR↗

arXiv2018

Image-Image Domain Adaptation with Preserved Self-Similarity and Domain-Dissimilarity for Person Re-identification

Person re-identification (re-ID) models trained on one domain often fail to generalize well to another. In our attempt, we present a "learning via translation" framework. In the baseline, we translate the labeled images from source to target domain in an unsupervised manner. We then train re-ID models with the translated images by supervised methods. Yet, being an essential part of this framework, unsupervised image-image translation suffers from the information loss of source-domain labels during translation. Our motivation is two-fold. First, for each image, the discriminative cues contained in its ID label should be maintained after translation. Second, given the fact that two domains have entirely different persons, a translated image should be dissimilar to any of the target IDs. To this end, we propose to preserve two types of unsupervised similarities, 1) self-similarity of an image before and after translation, and 2) domain-dissimilarity of a translated source image and a target image. Both constraints are implemented in the similarity preserving generative adversarial network (SPGAN) which consists of an Siamese network and a CycleGAN. Through domain adaptation experiment, we show that images generated by SPGAN are more suitable for domain adaptation and yield consistent and competitive re-ID accuracy on two large-scale datasets.

cs.CV↗

arXiv2018

The First Release of the AST3-1 Point Source Catalogue from Dome A, Antarctica

The three Antarctic Survey Telescopes (AST3) aim to carry out time domain imaging survey at Dome A, Antarctica. The first of the three telescopes (AST3-1) was successfully deployed on January 2012. AST3-1 is a 500\,mm aperture modified Schmidt telescope with a 680\,mm diameter primary mirror. AST3-1 is equipped with a SDSS $i$ filter and a 10k $\times$ 10k frame transfer CCD camera, reduced to 5k $\times$ 10k by electronic shuttering, resulting in a 4.3 deg$^2$ field-of-view. To verify the capability of AST3-1 for a variety of science goals, extensive commissioning was carried out between March and May 2012. The commissioning included a survey covering 2000 deg$^2$ as well as the entire Large and Small Magellanic Clouds. Frequent repeated images were made of the center of the Large Magellanic Cloud, a selected exoplanet transit field, and fields including some Wolf-Rayet stars. Here we present the data reduction and photometric measurements of the point sources observed by AST3-1. We have achieved a survey depth of 19.3\,mag in 60 s exposures with 5\,mmag precision in the light curves of bright stars. The facility achieves sub-mmag photometric precision under stable survey conditions, approaching its photon noise limit. These results demonstrate that AST3-1 at Dome A is extraordinarily competitive in time-domain astronomy, including both quick searches for faint transients and the detection of tiny transit signals.

astro-ph.IM↗

arXiv2018

Collective Entity Disambiguation with Structured Gradient Tree Boosting

We present a gradient-tree-boosting-based structured learning model for jointly disambiguating named entities in a document. Gradient tree boosting is a widely used machine learning algorithm that underlies many top-performing natural language processing systems. Surprisingly, most works limit the use of gradient tree boosting as a tool for regular classification or regression problems, despite the structured nature of language. To the best of our knowledge, our work is the first one that employs the structured gradient tree boosting (SGTB) algorithm for collective entity disambiguation. By defining global features over previous disambiguation decisions and jointly modeling them with local features, our system is able to produce globally optimized entity assignments for mentions in a document. Exact inference is prohibitively expensive for our globally normalized model. To solve this problem, we propose Bidirectional Beam Search with Gold path (BiBSG), an approximate inference algorithm that is a variant of the standard beam search algorithm. BiBSG makes use of global information from both past and future to perform better local search. Experiments on standard benchmark datasets show that SGTB significantly improves upon published results. Specifically, SGTB outperforms the previous state-of-the-art neural system by near 1\% absolute accuracy on the popular AIDA-CoNLL dataset.

cs.CL↗

arXiv2018

Adversarial Complementary Learning for Weakly Supervised Object Localization

In this work, we propose Adversarial Complementary Learning (ACoL) to automatically localize integral objects of semantic interest with weak supervision. We first mathematically prove that class localization maps can be obtained by directly selecting the class-specific feature maps of the last convolutional layer, which paves a simple way to identify object regions. We then present a simple network architecture including two parallel-classifiers for object localization. Specifically, we leverage one classification branch to dynamically localize some discriminative object regions during the forward pass. Although it is usually responsive to sparse parts of the target objects, this classifier can drive the counterpart classifier to discover new and complementary object regions by erasing its discovered regions from the feature maps. With such an adversarial learning, the two parallel-classifiers are forced to leverage complementary object regions for classification and can finally generate integral object localization together. The merits of ACoL are mainly two-fold: 1) it can be trained in an end-to-end manner; 2) dynamically erasing enables the counterpart classifier to discover complementary object regions more effectively. We demonstrate the superiority of our ACoL approach in a variety of experiments. In particular, the Top-1 localization error rate on the ILSVRC dataset is 45.14%, which is the new state-of-the-art.

cs.CV↗

arXiv2018

Camera Style Adaptation for Person Re-identification

Being a cross-camera retrieval task, person re-identification suffers from image style variations caused by different cameras. The art implicitly addresses this problem by learning a camera-invariant descriptor subspace. In this paper, we explicitly consider this challenge by introducing camera style (CamStyle) adaptation. CamStyle can serve as a data augmentation approach that smooths the camera style disparities. Specifically, with CycleGAN, labeled training images can be style-transferred to each camera, and, along with the original training samples, form the augmented training set. This method, while increasing data diversity against over-fitting, also incurs a considerable level of noise. In the effort to alleviate the impact of noise, the label smooth regularization (LSR) is adopted. The vanilla version of our method (without LSR) performs reasonably well on few-camera systems in which over-fitting often occurs. With LSR, we demonstrate consistent improvement in all systems regardless of the extent of over-fitting. We also report competitive accuracy compared with the state of the art.

cs.CV↗

arXiv2018

Occlusion Aware Unsupervised Learning of Optical Flow

It has been recently shown that a convolutional neural network can learn optical flow estimation with unsupervised learning. However, the performance of the unsupervised methods still has a relatively large gap compared to its supervised counterpart. Occlusion and large motion are some of the major factors that limit the current unsupervised learning of optical flow methods. In this work we introduce a new method which models occlusion explicitly and a new warping way that facilitates the learning of large motion. Our method shows promising results on Flying Chairs, MPI-Sintel and KITTI benchmark datasets. Especially on KITTI dataset where abundant unlabeled samples exist, our unsupervised method outperforms its counterpart trained with supervised learning.

cs.CV↗

arXiv2018

Differences in the gas and dust distribution in the transitional disk of a sun-like young star, PDS 70

We present ALMA 0.87 mm continuum, HCO+ J=4--3 emission line, and CO J=3--2 emission line data of the disk of material around the young, Sun-like star PDS 70. These data reveal the existence of a possible two component transitional disk system with a radial dust gap of 0."2 +/- 0."05, an azimuthal gap in the HCO+ J=4--3 moment zero map, as well as two bridge-like features in the gas data. Interestingly these features in the gas disk have no analogue in the dust disk making them of particular interest. We modeled the dust disk using the Monte Carlo radiative transfer code HOCHUNK3D (Whitney et al. 2013) using a two disk components. We find that there is a radial gap that extends from 15-60 au in all grain sizes which differs from previous work.

astro-ph.EP↗

arXiv2018

Peano Model for Planar Compacta and a Lemma by Beardon

It is known that, among all the monotone decompositions of a planar compact set K with Peano hyperspaces, there exists a unique one that is finer than all the others. We call it the "core decomposition" of K with Peano hyperspace. The resulted hyperspace under quotient topology will be referred to as the "Peano model" for K. We show that the core decomposition is independent of the embedding of K into the plane. Given a rational function f with degree at least 2 that is independent of K. A well known result by Beardon says that the pre-image for any element d of the core decomposition has finitely many components, each of which is mapped by f onto d. We show that those components belong to the core decomposition of L, the pre-image of K under the above mentioned rational map f. This provides an affirmative answer to Question 5.4 proposed by Curry (MR2642461) and extends earlier partial results by Blokh-Curry-Oversteegen (MR2737795 and MR3008890), when K is assumed to be unshielded and the rational map f is assumed to be a polynomial, under which K is completely invariant. The previous result is also connected with a well known Factor Theorem developed by Whyburn. We also introduce a lambda function from K to the set of non-negative integers, such that this function is constantly zero if and only if K is a Peano space. This function and its maximum are topologically invariant, while its level set at zero is of particular interest. For instance, when K is the Mandelbrot set we find close relations between the level set at zero and the hyperbolic components. Further discussions on the lambda function can be expected.

math.DS↗

arXiv2018

Style Aggregated Network for Facial Landmark Detection

Recent advances in facial landmark detection achieve success by learning discriminative features from rich deformation of face shapes and poses. Besides the variance of faces themselves, the intrinsic variance of image styles, e.g., grayscale vs. color images, light vs. dark, intense vs. dull, and so on, has constantly been overlooked. This issue becomes inevitable as increasing web images are collected from various sources for training neural networks. In this work, we propose a style-aggregated approach to deal with the large intrinsic variance of image styles for facial landmark detection. Our method transforms original face images to style-aggregated images by a generative adversarial module. The proposed scheme uses the style-aggregated image to maintain face images that are more robust to environmental changes. Then the original face images accompanying with style-aggregated ones play a duet to train a landmark detector which is complementary to each other. In this way, for each face, our method takes two images as input, i.e., one in its original style and the other in the aggregated style. In experiments, we observe that the large variance of image styles would degenerate the performance of facial landmark detectors. Moreover, we show the robustness of our method to the large variance of image styles by comparing to a variant of our approach, in which the generative adversarial module is removed, and no style-aggregated images are used. Our approach is demonstrated to perform well when compared with state-of-the-art algorithms on benchmark datasets AFLW and 300-W. Code is publicly available on GitHub: https://github.com/D-X-Y/SAN

cs.CV↗

arXiv2018

Strong light illumination on gain-switched semiconductor lasers helps the eavesdropper in practical quantum key distribution systems

The temperature of the semiconductor diode increases under strong light illumination whether thermoelectric cooler is installed or not, which changes the output wavelength of the laser (Lee M. S. et al., 2017). However, other characteristics also vary as temperature increases. These variations may help the eavesdropper in practical quantum key distribution systems. We study the effects of temperature increase on gain-switched semiconductor lasers by simulating temperature dependent rate equations. The results show that temperature increase may cause large intensity fluctuation, decrease the output intensity and lead the signal state and decoy state distinguishable. We also propose a modified photon number splitting attack by exploiting the effects of temperature increase. Countermeasures are also proposed.

quant-ph↗

arXiv2018

Multifrequency superscattering from subwavelength hyperbolic structures

Superscattering, i.e., a phenomenon of the scattering cross section from a subwavelength object exceeding the single-channel limit, has important prospects in enhanced sensing/spectroscopy, solar cells, and biomedical imaging. Superscattering can be typically constructed only at a single frequency regime, and depends critically on the inescapable material losses. Under such realistic conditions, superscattering has not been predicted nor observed to exist simultaneously at multiple frequency regimes. Here we introduce multifrequency superscattering in a subwavelength hyperbolic structure, which can be made from artificial metamaterials or from naturally-existing materials, such as hexagonal boron nitride (BN), and show the advantage of such hyperbolic materials for reducing structural complexity. The underlying mechanism is revealed to be the multimode resonances at multiple frequency regimes as appear in BN due to the peculiar dispersion of phonon-polaritons. Importantly, the multifrequency superscattering has a high tolerance to material losses and some structural variations, bringing the concept of multifrequency superscattering closer to useful and realistic conditions.

physics.optics↗

arXiv2018

Indexing of CNN Features for Large Scale Image Search

The convolutional neural network (CNN) features can give a good description of image content, which usually represent images with unique global vectors. Although they are compact compared to local descriptors, they still cannot efficiently deal with large-scale image retrieval due to the cost of the linear incremental computation and storage. To address this issue, we build a simple but effective indexing framework based on inverted table, which significantly decreases both the search time and memory usage. In addition, several strategies are fully investigated under an indexing framework to adapt it to CNN features and compensate for quantization errors. First, we use multiple assignment for the query and database images to increase the probability of relevant images' co-existing in the same Voronoi cells obtained via the clustering algorithm. Then, we introduce embedding codes to further improve precision by removing false matches during a search. We demonstrate that by using hashing schemes to calculate the embedding codes and by changing the ranking rule, indexing framework speeds can be greatly improved. Extensive experiments conducted on several unsupervised and supervised benchmarks support these results and the superiority of the proposed indexing framework. We also provide a fair comparison between the popular CNN features.

cs.CV↗