arXiv subjects

Liang Zheng

Publications and source records attributed to Liang Zheng.

At least 127 records · Page 7Linked to original sources

arXiv2021

Adaptive Affinity for Associations in Multi-Target Multi-Camera Tracking

Data associations in multi-target multi-camera tracking (MTMCT) usually estimate affinity directly from re-identification (re-ID) feature distances. However, we argue that it might not be the best choice given the difference in matching scopes between re-ID and MTMCT problems. Re-ID systems focus on global matching, which retrieves targets from all cameras and all times. In contrast, data association in tracking is a local matching problem, since its candidates only come from neighboring locations and time frames. In this paper, we design experiments to verify such misfit between global re-ID feature distances and local matching in tracking, and propose a simple yet effective approach to adapt affinity estimations to corresponding matching scopes in MTMCT. Instead of trying to deal with all appearance changes, we tailor the affinity metric to specialize in ones that might emerge during data associations. To this end, we introduce a new data sampling scheme with temporal windows originally used for data associations in tracking. Minimizing the mismatch, the adaptive affinity module brings significant improvements over global re-ID distance, and produces competitive performance on CityFlow and DukeMTMC datasets.

cs.CV↗

arXiv2021

Label-Free Model Evaluation with Semi-Structured Dataset Representations

Label-free model evaluation, or AutoEval, estimates model accuracy on unlabeled test sets, and is critical for understanding model behaviors in various unseen environments. In the absence of image labels, based on dataset representations, we estimate model performance for AutoEval with regression. On the one hand, image feature is a straightforward choice for such representations, but it hampers regression learning due to being unstructured (\ie no specific meanings for component at certain location) and of large-scale. On the other hand, previous methods adopt simple structured representations (like average confidence or average feature), but insufficient to capture the data characteristics given their limited dimensions. In this work, we take the best of both worlds and propose a new semi-structured dataset representation that is manageable for regression learning while containing rich information for AutoEval. Based on image features, we integrate distribution shapes, clusters, and representative samples for a semi-structured dataset representation. Besides the structured overall description with distribution shapes, the unstructured description with clusters and representative samples include additional fine-grained information facilitating the AutoEval task. On three existing datasets and 25 newly introduced ones, we experimentally show that the proposed representation achieves competitive results. Code and dataset are available at https://github.com/sxzrt/Semi-Structured-Dataset-Representations.

cs.CV↗

arXiv2021

Hierarchical Image Classification with A Literally Toy Dataset

Unsupervised domain adaptation (UDA) in image classification remains a big challenge. In existing UDA image dataset, classes are usually organized in a flattened way, where a plain classifier can be trained. Yet in some scenarios, the flat categories originate from some base classes. For example, buggies belong to the class bird. We define the classification task where classes have characteristics above and the flat classes and the base classes are organized hierarchically as hierarchical image classification. Intuitively, leveraging such hierarchical structure will benefit hierarchical image classification, e.g., two easily confusing classes may belong to entirely different base classes. In this paper, we improve the performance of classification by fusing features learned from a hierarchy of labels. Specifically, we train feature extractors supervised by hierarchical labels and with UDA technology, which will output multiple features for an input image. The features are subsequently concatenated to predict the finest-grained class. This study is conducted with a new dataset named Lego-15. Consisting of synthetic images and real images of the Lego bricks, the Lego-15 dataset contains 15 classes of bricks. Each class originates from a coarse-level label and a middle-level label. For example, class "85080" is associated with bricks (coarse) and bricks round (middle). In this dataset, we demonstrate that our method brings about consistent improvement over the baseline in UDA in hierarchical image classification. Extensive ablation and variant studies provide insights into the new dataset and the investigated algorithm.

cs.CV↗

arXiv2021

Investigation of exotic state X(3872) in $pp$ collisions at $\sqrt{s}=$7 and 13 TeV

We have used the dynamically constrained phase space coalescence model to study the production of the exotic state $X(3872)$ based on the hadronic final states generated by the parton and hadron cascade model (PACIAE) with $|y|<1$ and $p_T < 15.5$ GeV/c in $pp$ collisions at $\sqrt{s}=7$ and 13 TeV, respectively. Here the $X(3872)$ is assumed to consist of bound state $D\bar {D^*}$, which can form three possible structures for the tetraquark state, the nucleus-like state, and the molecular state. The yields of three different structures $X(3872)$ were predicted. The transverse momentum distribution and the rapidity distribution of three different structures $X(3872)$ are also presented. Sizable difference can be found in the transverse momentum and rapidity distributions for the three different $X(3872)$ structures.

hep-ph↗

arXiv2021

Synthetic Data Are as Good as the Real for Association Knowledge Learning in Multi-object Tracking

Association, aiming to link bounding boxes of the same identity in a video sequence, is a central component in multi-object tracking (MOT). To train association modules, e.g., parametric networks, real video data are usually used. However, annotating person tracks in consecutive video frames is expensive, and such real data, due to its inflexibility, offer us limited opportunities to evaluate the system performance w.r.t changing tracking scenarios. In this paper, we study whether 3D synthetic data can replace real-world videos for association training. Specifically, we introduce a large-scale synthetic data engine named MOTX, where the motion characteristics of cameras and objects are manually configured to be similar to those in real-world datasets. We show that compared with real data, association knowledge obtained from synthetic data can achieve very similar performance on real-world test sets without domain adaption techniques. Our intriguing observation is credited to two factors. First and foremost, 3D engines can well simulate motion factors such as camera movement, camera view and object movement, so that the simulated videos can provide association modules with effective motion features. Second, experimental results show that the appearance domain gap hardly harms the learning of association knowledge. In addition, the strong customization ability of MOTX allows us to quantitatively assess the impact of motion factors on MOT, which brings new insights to the community.

cs.CV↗

arXiv2021

Further developments of a multi-phase transport model for relativistic nuclear collisions

A multi-phase transport (AMPT) model was constructed as a self-contained kinetic theory-based description of relativistic nuclear collisions as it contains four main components: the fluctuating initial condition, a parton cascade, hadronization, and a hadron cascade. Here, we review the main developments after the first public release of the AMPT source code in 2004 and the corresponding publication that described the physics details of the model at that time. We also discuss possible directions for future developments of the AMPT model to better study the properties of the dense matter created in relativistic collisions of small or large systems.

nucl-th↗

arXiv2021

Ranking Models in Unlabeled New Environments

Consider a scenario where we are supplied with a number of ready-to-use models trained on a certain source domain and hope to directly apply the most appropriate ones to different target domains based on the models' relative performance. Ideally we should annotate a validation set for model performance assessment on each new target environment, but such annotations are often very expensive. Under this circumstance, we introduce the problem of ranking models in unlabeled new environments. For this problem, we propose to adopt a proxy dataset that 1) is fully labeled and 2) well reflects the true model rankings in a given target environment, and use the performance rankings on the proxy sets as surrogates. We first select labeled datasets as the proxy. Specifically, datasets that are more similar to the unlabeled target domain are found to better preserve the relative performance rankings. Motivated by this, we further propose to search the proxy set by sampling images from various datasets that have similar distributions as the target. We analyze the problem and its solutions on the person re-identification (re-ID) task, for which sufficient datasets are publicly available, and show that a carefully constructed proxy set effectively captures relative performance ranking in new environments. Code is available at \url{https://github.com/sxzrt/Proxy-Set}.

cs.CV↗

arXiv2021

Memory-Free Generative Replay For Class-Incremental Learning

Regularization-based methods are beneficial to alleviate the catastrophic forgetting problem in class-incremental learning. With the absence of old task images, they often assume that old knowledge is well preserved if the classifier produces similar output on new images. In this paper, we find that their effectiveness largely depends on the nature of old classes: they work well on classes that are easily distinguishable between each other but may fail on more fine-grained ones, e.g., boy and girl. In spirit, such methods project new data onto the feature space spanned by the weight vectors in the fully connected layer, corresponding to old classes. The resulting projections would be similar on fine-grained old classes, and as a consequence the new classifier will gradually lose the discriminative ability on these classes. To address this issue, we propose a memory-free generative replay strategy to preserve the fine-grained old classes characteristics by generating representative old images directly from the old classifier and combined with new data for new classifier training. To solve the homogenization problem of the generated samples, we also propose a diversity loss that maximizes Kullback Leibler (KL) divergence between generated samples. Our method is best complemented by prior regularization-based methods proved to be effective for easily distinguishable old classes. We validate the above design and insights on CUB-200-2011, Caltech-101, CIFAR-100 and Tiny ImageNet and show that our strategy outperforms existing memory-free methods with a clear margin. Code is available at https://github.com/xmengxin/MFGR

cs.CV↗

arXiv2021

Investigating high energy proton proton collisions with a multi-phase transport model approach based on PYTHIA8 initial conditions

The striking resemblance of high multiplicity proton-proton (pp) collisions at the LHC to heavy ion collisions challenges our conventional wisdom on the formation of the Quark-Gluon Plasma (QGP). A consistent explanation of the collectivity phenomena in pp will help us to understand the mechanism that leads to the QGP-like signals in small systems. In this study, we introduce a transport model approach connecting the initial conditions provided by PYTHIA8 with subsequent AMPT rescatterings to study the collective behavior in high energy pp collisions. The multiplicity dependence of light hadron productions from this model is in reasonable agreement with the pp $\sqrt{s}=13$ TeV experimental data. It is found in the comparisons that both the partonic and hadronic final state interactions are important for the generation of the radial flow feature of the pp transverse momentum spectra. The study also shows that the long range two particle azimuthal correlation in high multiplicity pp events is sensitive to the proton sub-nucleon spatial fluctuations.

hep-ph↗

arXiv2021

Multiview Detection with Shadow Transformer (and View-Coherent Data Augmentation)

Multiview detection incorporates multiple camera views to deal with occlusions, and its central problem is multiview aggregation. Given feature map projections from multiple views onto a common ground plane, the state-of-the-art method addresses this problem via convolution, which applies the same calculation regardless of object locations. However, such translation-invariant behaviors might not be the best choice, as object features undergo various projection distortions according to their positions and cameras. In this paper, we propose a novel multiview detector, MVDeTr, that adopts a newly introduced shadow transformer to aggregate multiview information. Unlike convolutions, shadow transformer attends differently at different positions and cameras to deal with various shadow-like distortions. We propose an effective training scheme that includes a new view-coherent data augmentation method, which applies random augmentations while maintaining multiview consistency. On two multiview detection benchmarks, we report new state-of-the-art accuracy with the proposed system. Code is available at https://github.com/hou-yz/MVDeTr.

cs.CV↗

arXiv2021

Collision system size scan for light (anti-)nuclei and (anti-)hypertriton production in high energy nuclear collisions

The production of light (anti-)nuclei and (anti-)hypertriton in a recent collsion system size scan program proposed for the STAR experiment at the Relativistic Heavy Ion Collider (RHIC) is investigated by using the dynamically constrained phase-space coalescence model and the parton and hadron cascade model. The collision system dependence of yield ratios for deuteron to proton, helium-3 to proton, and hypertriton to $Λ$-hyperon with the corresponding values for antiparticles is predicted. The work presents that for the yield ratios a significant difference exists between (hyper)nuclei and their anti-(hyper)nuclei. Besides, much more suppression for (anti-)hypernuclei than light (anti-)nuclei is present. We further investigate strangeness population factors $s_3$ as a function of atomic mass number $A$. Our present study can provide a reference for a upcoming collision system scan program at RHIC.

hep-ph↗

arXiv2021

Investigation of the background in coherent $J/ψ$ production at the EIC

Understanding various fundamental properties of nucleons and nuclei are among the most important scientific goals at the upcoming Electron-Ion Collider (EIC). With the unprecedented opportunity provided by the next-generation machine, the EIC might provide definitive answers to many standing puzzles and open questions in modern nuclear physics. Here we investigate one of the golden measurements proposed at the EIC, which is to obtain the spatial gluon density distribution within a lead ($Pb$) nucleus. The proposed experimental process is the exclusive $J/ψ$ vector-meson production off the $Pb$ nucleus - $e+Pb\rightarrow e'+J/ψ+Pb'$. The Fourier transformation of the momentum transfer $|t|$ distribution of the coherent diffraction is the transverse gluon spatial distribution. In order to measure it, the experiment has to overcome an overwhelmingly large background arising from the incoherent diffractive production, where the nucleus $Pb'$ mostly breaks up into fragments of particles in the far-forward direction close to the hadron-going beam rapidity. In this paper, we systematically study the rejection of incoherent $J/ψ$ production by vetoing products from these nuclear breakups - protons, neutrons, and photons, which is based on the BeAGLE event generator and the most up-to-date EIC Far-forward Interaction Region design. The achieved vetoing efficiency, the ratio between the number of vetoed events and total incoherent events, ranges from about 80% - 99% depending on $|t|$, which can resolve at least the first minimum of the coherent diffractive distribution based on the Sar$\it{t}$re model. Experimental and accelerator machine challenges as well as potential improvements are discussed.

nucl-ex↗

arXiv2021

Dual-Path Convolutional Image-Text Embeddings with Instance Loss

Matching images and sentences demands a fine understanding of both modalities. In this paper, we propose a new system to discriminatively embed the image and text to a shared visual-textual space. In this field, most existing works apply the ranking loss to pull the positive image / text pairs close and push the negative pairs apart from each other. However, directly deploying the ranking loss is hard for network learning, since it starts from the two heterogeneous features to build inter-modal relationship. To address this problem, we propose the instance loss which explicitly considers the intra-modal data distribution. It is based on an unsupervised assumption that each image / text group can be viewed as a class. So the network can learn the fine granularity from every image/text group. The experiment shows that the instance loss offers better weight initialization for the ranking loss, so that more discriminative embeddings can be learned. Besides, existing works usually apply the off-the-shelf features, i.e., word2vec and fixed visual feature. So in a minor contribution, this paper constructs an end-to-end dual-path convolutional network to learn the image and text representations. End-to-end learning allows the system to directly learn from the data and fully utilize the supervision. On two generic retrieval datasets (Flickr30k and MSCOCO), experiments demonstrate that our method yields competitive accuracy compared to state-of-the-art methods. Moreover, in language based person retrieval, we improve the state of the art by a large margin. The code has been made publicly available.

cs.CV↗

arXiv2021

Dynamical simulation on production of ${\rm W}^{\pm}$ and ${\rm Z}^{0}$ bosons in p-p, p-Pb (Pb-p), and Pb-Pb collisions at $\sqrt{s_{\rm NN}}=5.02$ TeV with PACIAE

In this paper, production of ${\rm W}^{\pm}$ and ${\rm Z}^{0}$ vector bosons in p-p, p-Pb (Pb-p), and Pb-Pb collisions at $\sqrt{s_{\rm NN}}=5.02$ TeV is dynamically simulated with a parton and hadron cascade model PACIAE. ALICE data of ${\rm Z}^{0}$ production is found to be reproduced fairly well. A prediction for ${\rm W}^{\pm}$ production is given in the same collision systems, at the same energy and at the same energy. An interesting isospin-effect is observed in the sign-change of $μ^{\pm}$ charge asymmetry in pp, pn, np, and nn collisions and in minimum bias p-Pb, Pb-p and Pb-Pb collisions at $\sqrt{s_{\rm NN}}=5.02$ TeV, respectively.

nucl-th↗

arXiv2021

Invertible Attention

Attention has been proved to be an efficient mechanism to capture long-range dependencies. However, so far it has not been deployed in invertible networks. This is due to the fact that in order to make a network invertible, every component within the network needs to be a bijective transformation, but a normal attention block is not. In this paper, we propose invertible attention that can be plugged into existing invertible models. We mathematically and experimentally prove that the invertibility of an attention model can be achieved by carefully constraining its Lipschitz constant. We validate the invertibility of our invertible attention on image reconstruction task with 3 popular datasets: CIFAR-10, SVHN, and CelebA. We also show that our invertible attention achieves similar performance in comparison with normal non-invertible attention on dense prediction tasks. The code is available at https://github.com/Schwartz-Zha/InvertibleAttention

cs.CV↗

arXiv2021

Combining Semantic Guidance and Deep Reinforcement Learning For Generating Human Level Paintings

Generation of stroke-based non-photorealistic imagery, is an important problem in the computer vision community. As an endeavor in this direction, substantial recent research efforts have been focused on teaching machines "how to paint", in a manner similar to a human painter. However, the applicability of previous methods has been limited to datasets with little variation in position, scale and saliency of the foreground object. As a consequence, we find that these methods struggle to cover the granularity and diversity possessed by real world images. To this end, we propose a Semantic Guidance pipeline with 1) a bi-level painting procedure for learning the distinction between foreground and background brush strokes at training time. 2) We also introduce invariance to the position and scale of the foreground object through a neural alignment model, which combines object localization and spatial transformer networks in an end to end manner, to zoom into a particular semantic instance. 3) The distinguishing features of the in-focus object are then amplified by maximizing a novel guided backpropagation based focus reward. The proposed agent does not require any supervision on human stroke-data and successfully handles variations in foreground object attributes, thus, producing much higher quality canvases for the CUB-200 Birds and Stanford Cars-196 datasets. Finally, we demonstrate the further efficacy of our method on complex datasets with multiple foreground object instances by evaluating an extension of our method on the challenging Virtual-KITTI dataset. Source code and models are available at https://github.com/1jsingh/semantic-guidance.

cs.CV↗

arXiv2021

What Does Rotation Prediction Tell Us about Classifier Accuracy under Varying Testing Environments?

Understanding classifier decision under novel environments is central to the community, and a common practice is evaluating it on labeled test sets. However, in real-world testing, image annotations are difficult and expensive to obtain, especially when the test environment is changing. A natural question then arises: given a trained classifier, can we evaluate its accuracy on varying unlabeled test sets? In this work, we train semantic classification and rotation prediction in a multi-task way. On a series of datasets, we report an interesting finding, i.e., the semantic classification accuracy exhibits a strong linear relationship with the accuracy of the rotation prediction task (Pearson's Correlation r > 0.88). This finding allows us to utilize linear regression to estimate classifier performance from the accuracy of rotation prediction which can be obtained on the test set through the freely generated rotation labels.

cs.CV↗

arXiv2021

Are Labels Always Necessary for Classifier Accuracy Evaluation?

To calculate the model accuracy on a computer vision task, e.g., object recognition, we usually require a test set composing of test samples and their ground truth labels. Whilst standard usage cases satisfy this requirement, many real-world scenarios involve unlabeled test data, rendering common model evaluation methods infeasible. We investigate this important and under-explored problem, Automatic model Evaluation (AutoEval). Specifically, given a labeled training set and a classifier, we aim to estimate the classification accuracy on unlabeled test datasets. We construct a meta-dataset: a dataset comprised of datasets generated from the original images via various transformations such as rotation, background substitution, foreground scaling, etc. As the classification accuracy of the model on each sample (dataset) is known from the original dataset labels, our task can be solved via regression. Using the feature statistics to represent the distribution of a sample dataset, we can train regression models (e.g., a regression neural network) to predict model performance. Using synthetic meta-dataset and real-world datasets in training and testing, respectively, we report a reasonable and promising prediction of the model accuracy. We also provide insights into the application scope, limitation, and potential future direction of AutoEval.

cs.CV↗