A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2022; you can also visit the original URL.
The file type is application/pdf.
Filters
BViT: Broad Attention based Vision Transformer
[article]
2022
arXiv
pre-print
Preserved Fulltext
Li, Y. Chen, W. Li, Z. Ding and D. ...
arXiv:2202.06268v1
fatcat:kef3cztfkzbyrkb2lqzry2zwry
Learning Compact Reward for Image Captioning
[article]
2020
arXiv
pre-print
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2020; you can also visit the original URL.
The file type is application/pdf.
However, due to the reward hacking problem, RL maximizes the reward in an unintended way and fails to produce human-like descriptions (Li et al., 2019a) . ...
Although AIRL has been utilized to solve problems in other fields (Wang et al., 2018; Li et al., 2019b; Shi et al., 2018) , we are the first to make algorithmic improvements to AIRL such that the model ...
arXiv:2003.10925v1
fatcat:se65gwk7fvae3kfpo6jyxvvase
Combined Use of eDNA Metabarcoding and Bottom Trawling for the Assessment of Fish Biodiversity in the Zhoushan Sea
2022
Frontiers in Marine Science
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2022; you can also visit the original URL.
The file type is application/pdf.
With the increase in fishing intensity and the intensification of marine pollution, the fishery resources in the Zhoushan Sea are seriously degraded, and the difficulty of censusing fish diversity hampers effective management in marine fishes. Environmental DNA metabarcoding and bottom trawl methods were used to determine the ability of the methods to distinguish fish assemblages in the Zhoushan Sea. The species composition and diversity of the Zhoushan Sea were assessed via high-throughput
doi:10.3389/fmars.2021.809703
doaj:b0b99ca415bf431da9d8cf9b86717247
fatcat:2j7p4k65ajaelcqh4jinngjclq
more »
... encing analysis of eDNA coupled with bottom trawl fishery survey data, after which the two methods were compared. eDNA screening identified 38.2% more fish species than bottom trawls. Combining these two methods, 33 orders, 65 families, and 130 species of fishes were identified. Perciformes and Clupeiformes, the most abundant orders in the catch, represented 31.5 and 10.0% of the total fish abundance, respectively. The results of ANOSIM and redundancy analyses indicated that the fish community structure varied significantly between summer and winter, however depth and temperature being the main environmental factors influencing fish distribution. The biodiversity index was higher in summer than in winter. Thus, our work provides more detailed seasonal data on biodiversity in the Zhoushan Sea, which is essential for the long-term management and conservation of coastal biodiversity. Compared with traditional survey methods, eDNA determination is highly sensitive, accurate, cost-efficient, and suitable for fish diversity studies in relevant sea areas. Although this approach cannot completely replace traditional methods, our findings demonstrate that it provides a reliable complementary method for assessing fish diversity in marine ecosystems.
Predicate correlation learning for scene graph generation
[article]
2021
arXiv
pre-print
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2021; you can also visit the original URL.
The file type is application/pdf.
Based on the conclusion of , Li et al. (Li et al., 2020) enhanced samples of hard classes by group softmax. ...
., 2018b; Li et al., 2018; Mi and Chen, 2020; Xu et al., 2017; Ren et al., 2020; Hung et al., 2020; Guo et al., 2021 ) focused on better modeling the interactions among objects. Xu et al. ...
arXiv:2107.02713v1
fatcat:amxwiwrcn5fbrdy6bfxg6mmnpm
Heterogeneous image transformation
2013
Pattern Recognition Letters
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2019; you can also visit the original URL.
The file type is application/pdf.
Li et al. (2006) extended the eigen-transform to a hybrid space consisting of sketch space and photo space. ...
applications, there are diverse imaging modes and correspondingly different image modalities for different purpose, such as near infrared (NIR) image applied for illumination invariant face recognition (Li ...
doi:10.1016/j.patrec.2012.04.005
fatcat:vkqvusm2t5auxfzlbxqjgbvrnu
Multi-Margin based Decorrelation Learning for Heterogeneous Face Recognition
[article]
2020
arXiv
pre-print
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2020; you can also visit the original URL.
The file type is application/pdf.
Heterogeneous face recognition (HFR) refers to matching face images acquired from different domains with wide applications in security scenarios. This paper presents a deep neural network approach namely Multi-Margin based Decorrelation Learning (MMDL) to extract decorrelation representations in a hyperspherical space for cross-domain face images. The proposed framework can be divided into two components: heterogeneous representation network and decorrelation representation learning. First, we
arXiv:2005.11945v1
fatcat:hwdx23hbxrbmri3pfne5spmoyy
more »
... mploy a large scale of accessible visual face images to train heterogeneous representation network. The decorrelation layer projects the output of the first component into decorrelation latent subspace and obtains decorrelation representation. In addition, we design a multi-margin loss (MML), which consists of quadruplet margin loss (QML) and heterogeneous angular margin loss (HAML), to constrain the proposed framework. Experimental results on two challenging heterogeneous face databases show that our approach achieves superior performance on both verification and recognition tasks, comparing with state-of-the-art methods.
Facial Attribute Capsules for Noise Face Super Resolution
[article]
2020
arXiv
pre-print
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2020; you can also visit the original URL.
The file type is application/pdf.
Existing face super-resolution (SR) methods mainly assume the input image to be noise-free. Their performance degrades drastically when applied to real-world scenarios where the input image is always contaminated by noise. In this paper, we propose a Facial Attribute Capsules Network (FACN) to deal with the problem of high-scale super-resolution of noisy face image. Capsule is a group of neurons whose activity vector models different properties of the same entity. Inspired by the concept of
arXiv:2002.06518v1
fatcat:zs6355bui5aqpfric55kmv5rry
more »
... ule, we propose an integrated representation model of facial information, which named Facial Attribute Capsule (FAC). In the SR processing, we first generated a group of FACs from the input LR face, and then reconstructed the HR face from this group of FACs. Aiming to effectively improve the robustness of FAC to noise, we generate FAC in semantic, probabilistic and facial attributes manners by means of integrated learning strategy. Each FAC can be divided into two sub-capsules: Semantic Capsule (SC) and Probabilistic Capsule (PC). Them describe an explicit facial attribute in detail from two aspects of semantic representation and probability distribution. The group of FACs model an image as a combination of facial attribute information in the semantic space and probabilistic space by an attribute-disentangling way. The diverse FACs could better combine the face prior information to generate the face images with fine-grained semantic attributes. Extensive benchmark experiments show that our method achieves superior hallucination results and outperforms state-of-the-art for very low resolution (LR) noise face image super resolution.
Video Face Super-Resolution with Motion-Adaptive Feedback Cell
[article]
2020
arXiv
pre-print
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2020; you can also visit the original URL.
The file type is application/pdf.
Li et.al [32] optimized the structure of 3D convolution and proposes a fast VSR method. ...
arXiv:2002.06378v1
fatcat:sf2ioi4sjzbppodz5723num53m
Searching Action Proposals via Spatial Actionness Estimation and Temporal Path Inference and Tracking
[article]
2016
arXiv
pre-print
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2020; you can also visit the original URL.
The file type is application/pdf.
n
g
F
e
n
c
in
g
R
o
p
e
C
li
m
b
in
g
Ic
e
D
a
n
c
in
g
W
a
lk
in
g
W
it
h
D
o
g
F
lo
o
rG
y
m
n
a
s
ti
c
s
B
ik
in
g
H
o
rs
e
R
id ...
in
g
T
ra
m
p
o
li
n
e
J
u
m
p
in
g
S
k
a
te
B
o
a
rd
in
g
G
o
lf
S
w
in
g
S
a
ls
a
S
p
in
Table 1 . 1 Quantitative performance comparison of the action ...
arXiv:1608.06495v1
fatcat:dv6jnd2nq5gvzjdg74ewnbsqpa
Facial Feature Point Detection: A Comprehensive Survey
[article]
2014
arXiv
pre-print
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2020; you can also visit the original URL.
The file type is application/pdf.
Li et al. (Li et al, 2009 , 2011 thus presented a robust shape model exploring random sample consensus (Fischler and Bolles, 1981) . ...
better performance; (2) the generic shape model is far easier to build than the generic texture model; (3) the origin of the idea that fitting the generic AAM is far harder than the person specific AAM lies ...
arXiv:1410.1037v1
fatcat:ye2nza4tabdehpk6rr62cvwute
OH Evolution in Molecular Clouds
[article]
2020
arXiv
pre-print
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2020; you can also visit the original URL.
The file type is application/pdf.
We adopted the same analysis method of cloud age in Goldsmith & Li (2005) . ...
-Y., Zuo, P., Li, D., et al. 2020 Astronomy and Astrophysics, 20, 077 Tang, N., Li, D., Heiles, C., et al. 2017 , ApJ, 839, 8 Turner, B. E. 1973 , ApJ, 186, 357 Turner, B. E. & Heiles, C. ...
arXiv:2010.09264v1
fatcat:hvnhldgon5ahhlxxene56k2cme
Non-rigid 3D shape retrieval based on multi-view metric learning
[article]
2019
arXiv
pre-print
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2019; you can also visit the original URL.
The file type is application/pdf.
This study presents a novel multi-view metric learning algorithm, which aims to improve 3D non-rigid shape retrieval. With the development of non-rigid 3D shape analysis, there exist many shape descriptors. The intrinsic descriptors can be explored to construct various intrinsic representations for non-rigid 3D shape retrieval task. The different intrinsic representations (features) focus on different geometric properties to describe the same 3D shape, which makes the representations are
arXiv:1904.00765v1
fatcat:x6brnuwcgncvjdbddktporz6hm
more »
... . Therefore, it is possible and necessary to learn multiple metrics for different representations jointly. We propose an effective multi-view metric learning algorithm by extending the Marginal Fisher Analysis (MFA) into the multi-view domain, and exploring Hilbert-Schmidt Independence Criteria (HSCI) as a diversity term to jointly learning the new metrics. The different classes can be separated by MFA in our method. Meanwhile, HSCI is exploited to make the multiple representations to be consensus. The learned metrics can reduce the redundancy between the multiple representations, and improve the accuracy of the retrieval results. Experiments are performed on SHREC'10 benchmarks, and the results show that the proposed method outperforms the state-of-the-art non-rigid 3D shape retrieval methods.
Gene analysis of seven cases of primary immunodeficiency
2020
Translational Pediatrics
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2020; you can also visit the original URL.
The file type is application/pdf.
Primary immune deficiency diseases (PID) are a group of potentially serious disorders in which inherited defects in the immune system lead to increased infections. This paper explores the clinical characteristics and pathogenic gene mutation of PID. The clinical data, clinical manifestations, and gene sequencing results of seven children were analyzed. Among the seven children, six were male, and one was female, aged from 4 months to 13 years old. All of them had a history of repeated infection
doi:10.21037/tp.2020.03.07
pmid:32477911
pmcid:PMC7237979
fatcat:g3bcps5p6jeyxor2plnx7nzcym
more »
... and pneumonia. High throughput sequencing (NGS) showed that the BTK gene of case 1 had c.1921c > t mutation; the BTK gene of case 2 had c.906-908del splice site mutation; the BTK gene of case 3 had c.718delg mutation; the cybb gene of case 4 had c.469c > t mutation; the IL2RG gene of case 5 had c.202g > A mutation; the STAT1 gene of case 6 had c.854a > G mutation; the case 7 had c.718delg mutation. There was c.1154c > t mutation in the STAT1 gene. Cases 1, 3, 6 and 7 were new mutations, and cases 2, 4, and 5 were inherited from mothers. In clinical cases of children with recurrent infection, the immunologic index is abnormal, so we need to be highly aware of the possibility of PID, and timely high-throughput sequencing is helpful for the diagnosis.
Field Emission and Emission-Stimulated Desorption of ZnO Nanomaterials
2018
Applied Sciences
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2019; you can also visit the original URL.
The file type is application/pdf.
Author Contributions: Nannan Li performed the fabrication and field emission test experiments, analyzed the data, and wrote an article; Xiaozhao Li help to analyze the data and participated in the discussion ...
doi:10.3390/app8030382
fatcat:25jza6ofi5efxd25neaixo4uda
ABCP: Automatic Block-wise and Channel-wise Network Pruning via Joint Search
[article]
2021
arXiv
pre-print
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2021; you can also visit the original URL.
The file type is application/pdf.
Haoran Li, Yaran Chen, Zixiang Ding, Nannan Li, Mingjun Ma and Dongbin Zhao are with the State Key Laboratory of Management and Control for Complex Systems, Institute of Automation, Chinese Academy of ...
(ρ i ) (9) π(a li |s; θ µ li , θ ρ li ) ∼ N (μ i ,σ i 2 ) (10) a li ∼ π(a li ∈ C|s; θ µ li , θ ρ li ) (11) where µ i and ρ i in (7) and (8) are the two FC layers in the ith LSTM cell that estimate ...
arXiv:2110.03858v1
fatcat:qvkpuiqplvbujie6a2rjl6otta
« Previous
Showing results 1 — 15 out of 1,253 results