Source author record

Xiaojia Chen

Xiaojia Chen appears in the imported research catalog. Authorship, coauthor and topic links are available while profile ownership is still unclaimed.

ResearcherUnclaimed source record

Computer Vision cond-mat.supr-con cond-mat.mtrl-sci cond-mat.str-el physics.comp-ph

Catalog footprint

What is connected

5works

5topics

4close collaborators

Actions

Connect this record

Open graph Browse works

Inspect adjacent papers, topics, institutions and collaborators without losing the researcher page.

Building this map preview

BZPEER is loading the nearby papers, people, topics and institutions for this page.

preprint2022arXiv

KE-RCNN: Unifying Knowledge based Reasoning into Part-level Attribute Parsing

Part-level attribute parsing is a fundamental but challenging task, which requires the region-level visual understanding to provide explainable details of body parts. Most existing approaches address this problem by adding a regional convolutional neural network (RCNN) with an attribute prediction head to a two-stage detector, in which attributes of body parts are identified from local-wise part boxes. However, local-wise part boxes with limit visual clues (i.e., part appearance only) lead to unsatisfying parsing results, since attributes of body parts are highly dependent on comprehensive relations among them. In this article, we propose a Knowledge Embedded RCNN (KE-RCNN) to identify attributes by leveraging rich knowledges, including implicit knowledge (e.g., the attribute ``above-the-hip'' for a shirt requires visual/geometry relations of shirt-hip) and explicit knowledge (e.g., the part of ``shorts'' cannot have the attribute of ``hoodie'' or ``lining''). Specifically, the KE-RCNN consists of two novel components, i.e., Implicit Knowledge based Encoder (IK-En) and Explicit Knowledge based Decoder (EK-De). The former is designed to enhance part-level representation by encoding part-part relational contexts into part boxes, and the latter one is proposed to decode attributes with a guidance of prior knowledge about \textit{part-attribute} relations. In this way, the KE-RCNN is plug-and-play, which can be integrated into any two-stage detectors, e.g., Attribute-RCNN, Cascade-RCNN, HRNet based RCNN and SwinTransformer based RCNN. Extensive experiments conducted on two challenging benchmarks, e.g., Fashionpedia and Kinetics-TPS, demonstrate the effectiveness and generalizability of the KE-RCNN. In particular, it achieves higher improvements over all existing methods, reaching around 3% of AP on Fashionpedia and around 4% of Acc on Kinetics-TPS.

preprint2022arXiv

RepParser: End-to-End Multiple Human Parsing with Representative Parts

Existing methods of multiple human parsing usually adopt a two-stage strategy (typically top-down and bottom-up), which suffers from either strong dependence on prior detection or highly computational redundancy during post-grouping. In this work, we present an end-to-end multiple human parsing framework using representative parts, termed RepParser. Different from mainstream methods, RepParser solves the multiple human parsing in a new single-stage manner without resorting to person detection or post-grouping.To this end, RepParser decouples the parsing pipeline into instance-aware kernel generation and part-aware human parsing, which are responsible for instance separation and instance-specific part segmentation, respectively. In particular, we empower the parsing pipeline by representative parts, since they are characterized by instance-aware keypoints and can be utilized to dynamically parse each person instance. Specifically, representative parts are obtained by jointly localizing centers of instances and estimating keypoints of body part regions. After that, we dynamically predict instance-aware convolution kernels through representative parts, thus encoding person-part context into each kernel responsible for casting an image feature as an instance-specific representation.Furthermore, a multi-branch structure is adopted to divide each instance-specific representation into several part-aware representations for separate part segmentation.In this way, RepParser accordingly focuses on person instances with the guidance of representative parts and directly outputs parsing results for each person instance, thus eliminating the requirement of the prior detection or post-grouping.Extensive experiments on two challenging benchmarks demonstrate that our proposed RepParser is a simple yet effective framework and achieves very competitive performance.

preprint2016arXiv

Magnetic ground state of FeSe

Elucidating the nature of the magnetism of a high-temperature superconductor is crucial for establishing its pairing mechanism. The parent compounds of the cuprate and iron-pnictide superconductors exhibit Néel and stripe magnetic order, respectively. However, FeSe, the structurally simplest iron-based superconductor, shows nematic order (Ts = 90 K), but not magnetic order in the parent phase, and its magnetic ground state is intensely debated. Here, we report inelastic neutron-scattering experiments that reveal both stripe and Néel spin fluctuations over a wide energy range at 110 K. On entering the nematic phase, a substantial amount of spectral weight is transferred from the Néel to the stripe spin fluctuations. Moreover, the total fluctuating magnetic moment of FeSe is ~ 60% larger than that in the iron pnictide BaFe2As2. Our results suggest that FeSe is a novel S = 1 nematic quantum-disordered paramagnet interpolating between the Néel and stripe magnetic instabilities.

preprint2015arXiv

Geometrical and electronic structures of tripotassium-doped hydrocarbon superconductors: Density functional calculations

A systemically theoretical study has been presented to explored the crystal structures and electronic characteristics of polycyclic aromatic hydrocarbons (PAHs), such as solid phenanthrene, picene, 1,2;8,9-dibenzopentacene, and 7-phenacenes, since these PAHs exhibited the superconductivity when potassium doping into. For tripotassium-doped phenanthrene and picene, we demonstrate the K atomic positions to fit the experimental lattice parameters, and analyze the distinction between the stablest configuration and the fitted experimental one. Based on the first-principles calculations, for the first time, we predict the possible crystal configurations of pristine and tripotassium-doped 1,2;8,9-dibenzopentacene and 7-phenacenes, respectively. For these four PAHs, the electronic structures after doping are investigated in details. The results show that the electronic characters near the Fermi level are high sensitive to structure. Because of the change of the benzene rings arrangement, the 1,2;8,9-dibenzopentacene exhibits visibly different band structures from other three PAHs. In these metallic PAHs, two bands cross the Fermi level which results in the complicated multiband feature of Fermi surfaces. Fascinatingly, we find that the electronic states of potassium contribute to the Fermi surfaces especially for K-3$d$ electrons, which improves a way to understand this superconductivity. As a result, we suggest that the rigid-band picture is invalidated due to the hybridization between K atoms and PAH molecules as well as the rearrangement and distortion of PAH molecules.

preprint2012arXiv

Pressure-Driven Quantum Criticality in An Iron-Selenide Superconductor

The discovery of superconductivity of about 30 K in iron selenides with very large magnetic moments simulates the examination of completing orders. Here we report a finding of pressure- induced suppression of the superconducting transition temperature Tc and enhancement of the temperature of the resistance hump TH through charge transfer between two iron sites with different occupancies. The activation energy for the electric transport of the high-temperature resistance is observed to go to zero at a critical pressure of 8.7 GPa, at which superconductivity tends to disappear and the semiconductor-to-metal transition takes place. Beyond the critical point, the resistance exhibits a metallic behavior over the whole temperature range studied. All these features indicate the existence of quantum criticality in iron-selenide superconductors.