Filters








4,699 Hits in 2.7 sec

Automatic Spatially-aware Fashion Concept Discovery [article]

Xintong Han, Zuxuan Wu, Phoenix X. Huang, Xiao Zhang, Menglong Zhu, Yuan Li, Yang Zhao, Larry S. Davis
2017 arXiv   pre-print
Huang et al. [8] built treestructured layers for all attribute categories to form a semantic representation for clothing images. Veit et al.  ...  If we plug f k into the cosine distance we obtain: d(x, W a ) = X m W a m x m = X m W a m X k W I m,k f k = X m W a m X k W I m,k X i,j q k (i, j) = X i,j X m W a m X k W I m,k q k (i, j) (2) where W a  ... 
arXiv:1708.01311v1 fatcat:exsjmxh7vzfidfea62igpunyfe

Implementation of an Automated Learning System for Non-experts [article]

Phoenix X. Huang, Zhiwei Zhao, Chao Liu, Jingyi Liu, Wenze Hu, Xiaoyu Wang
2022 arXiv   pre-print
Create a message subscription group x for the Redis stream, with only one consumer 2.  ...  It is noted that while the high-level concept design of YMIR is introduced in [Huang et al., 2021] , this paper emphasis on the engineering design principles as well as implementation details.  ... 
arXiv:2203.15784v1 fatcat:2gj6rg6sovgfbbenqzafnywv5q

Underwater Live Fish Recognition Using a Balance-Guaranteed Optimized Tree [chapter]

Phoenix X. Huang, Bastiaan J. Boom, Robert B. Fisher
2013 Lecture Notes in Computer Science  
u (u, σ)Y uu (u, σ) − X uu (u, σ)Y u (u, σ) (X u (u, σ) 2 + Y u (u, σ) 2 )) 3 2 (4) where X u (u, σ)/X uu (u, σ) and Y u (u, σ)/Y uu (u, σ) are the first and the second derivative of X(u, σ) and Y (u,  ...  σ), respectively; X(u, σ) and Y (u, σ) are the convolution result of 1-D Gaussian kernel function g(u, σ) with fish boundary coordinates x(u) and y(u).  ... 
doi:10.1007/978-3-642-37331-2_32 fatcat:cij4sanu6fgzzp4fqfcfj7to5i

Hierarchical classification with reject option for live fish recognition

Phoenix X. Huang, Bastiaan J. Boom, Robert B. Fisher
2014 Machine Vision and Applications  
x i .  ...  Rejection is based on the posterior probability for the observation evidence X with the predicted class C i (using Bayes' Rule): p(C i | X) = p(C i )p(X | C i ) p(X) = p(C i )p(X | C i ) j p(C j )p(X |  ... 
doi:10.1007/s00138-014-0641-2 fatcat:x7lqyu7w4jcqvdqrt5sofzrkhe

YMIR: A Rapid Data-centric Development Platform for Vision Applications [article]

Phoenix X. Huang, Wenze Hu, William Brendel, Manmohan Chandraker, Li-Jia Li, Xiaoyu Wang
2021 arXiv   pre-print
X.  ...  YMIR: A Rapid Data-centric Development Platform for Vision Applications Phoenix  ... 
arXiv:2111.10046v2 fatcat:3k5bmhrl7ra37gqfdrlyxxnz6q

Applying semi-synchronised task farming to large-scale computer vision problems

Steven McDonagh, Cigdem Beyan, Phoenix X Huang, Robert B Fisher
2014 The international journal of high performance computing applications  
The error function erf is defined as: 306 erf(x) = 2 √ π x 0 e −t 2 dt Then the complementary error function, denoted erfc and its inverse erfc −1 are defined as: erfc(x) = 1 − erf(x) = 2 √ π ∞ x e −t  ...  2 dt erfc −1 (1 − x) = erf −1 (x) The model that empirically fits the simulation for mean task length w µ , with standard deviation σ 307 distributing N s tasks in parallel, lets us predict the maximum  ... 
doi:10.1177/1094342014532965 fatcat:3bzpwun6ancgjjdaehbzriwjwi

GMM improves the reject option in hierarchical classification for fish recognition

Phoenix X. Huang, Bastiaan J. Boom, Robert B. Fisher
2014 IEEE Winter Conference on Applications of Computer Vision  
More specifically, the rejection uses the posterior probability for the predicted class C i given evidence X: p(C i | X) = p(C i )p(X | C i ) p(X) = p(C i )p(X | C i ) j p(C j )p(X | C j ) (2) where the  ...  ) D 2 | Σ i | 1 2 exp{− 1 2 (x−µ i ) Σ −1 i (x−µ i )} (1) θ is the parameters of infinite mixture model, including ω i and µ i and Σ i , g(x | µ i , Σ i ) is the component Gaussian density, while each  ... 
doi:10.1109/wacv.2014.6836076 dblp:conf/wacv/HuangBF14 fatcat:f3ogewugtzakvi5qco2qxhtgyq

A research tool for long-term and continuous analysis of fish assemblage in coral-reefs using underwater camera footage

Bastiaan J. Boom, Jiyin He, Simone Palazzo, Phoenix X. Huang, Cigdem Beyan, Hsiu-Mei Chou, Fang-Pang Lin, Concetto Spampinato, Robert B. Fisher
2014 Ecological Informatics  
Each vector contains the pixel's (x, y) coordinates, RGB values, hue value and the mean of the grayscale histogram in a 5 × 5 window.  ...  For instance, the distribution of fish counts over different species recorded by camera X may be different from that recorded by camera Y, as the cameras are located in different areas.  ... 
doi:10.1016/j.ecoinf.2013.10.006 fatcat:my7dwck43ng7fgt6qed7kuonni

Approximate nearest neighbor search to support manual image annotation of large domain-specific datasets

Bastiaan J. Boom, Phoenix X. Huang, Robert B. Fisher
2013 Proceedings of the International Workshop on Video and Image Ground Truth in Computer Vision Applications - VIGTA '13  
Given a joint distribution p(x, y) on the "model" space X and the "feature" space Y , find a clusteringX that minimizes the information loss I(X; Y ) − I(X; Y ), where I(X; Y ) is the mutual information  ...  between X and Y .  ... 
doi:10.1145/2501105.2501112 dblp:conf/vigta/BoomHF13 fatcat:44dokr6bdzecxfk377xow7arw4

Density Analysis and Carbon Age Estimation of Taiwan Cypress Burl Wood Figure

Wei–Chih HUANG, Ming–Chun JANE, Noboru FUJIMOTO, Han Chien LIN
2021 Journal of the Faculty of Agriculture Kyushu University  
reports (Huang et al., 2020) .  ...  Soft X-ray density analysis of phoenix tail-liked figure of Chamaecyparis formosensis specimen Table 1 and punctate inclusion with the part of wood ray pattern, such as crossover region.  ... 
doi:10.5109/4363554 fatcat:nindaepu7jfpzefnr7vsyvmjym

Regional-scale transport of air pollutants: impacts of Southern California emissions on Phoenix ground-level ozone concentrations

J. Li, M. Georgescu, P. Hyde, A. Mahalov, M. Moustaoui
2015 Atmospheric Chemistry and Physics  
SoCal contributions to DMA8 [O<sub>3</sub>] for the Phoenix metropolitan area range from a few ppbv to over 30 ppbv (10–30 % relative to Control experiments).  ...  Based on the USEPA NEI05, results for the selected events indicate the impacts of AZ emissions are dominant on daily maximum 8 h average (DMA8) [O<sub>3</sub>] in Phoenix.  ...  Following Huang et al. (2013), the contribution of SoCal to [O 3 ] in the Phoenix area is the differ- ence between the CTRL and noCA experiments.  ... 
doi:10.5194/acp-15-9345-2015 fatcat:dwefmyhvujh3zfvkmzqzqlycty

Regional-scale transport of air pollutants: impacts of southern California emissions on Phoenix ground-level ozone concentrations

J. Li, M. Georgescu, P. Hyde, A. Mahalov, M. Moustaoui
2015 Atmospheric Chemistry and Physics Discussions  
Results for the selected events indicate the impacts of AZ emissions are dominant on daily maximum 8 h average (DMA8) [O<sub>3</sub>] in Phoenix.  ...  SoCal contributions to DMA8 [O<sub>3</sub>] for the Phoenix metropolitan area range from a few ppbv to over 30 ppbv (10–30% relative to Control experiments).  ...  Following Huang et al. (2013), the contribution of SoCal to [O 3 ] in the Phoenix area is the differ- ence between the CTRL and noCA experiments.  ... 
doi:10.5194/acpd-15-8361-2015 fatcat:lwekpgf7krfxxlgo75sy7rlvwm

Spatial-Temporal Multi-Cue Network for Continuous Sign Language Recognition [article]

Hao Zhou, Wengang Zhou, Yun Zhou, Houqiang Li
2020 arXiv   pre-print
To validate the effectiveness, we perform experiments on three large-scale CSLR benchmarks: PHOENIX-2014, CSL and PHOENIX-2014-T.  ...  We evaluate our method on three datasets, including PHOENIX-2014 (Koller, Forster, and Ney 2015) , CSL (Huang et al. 2018; and PHOENIX-2014 -T (Cihan Camgoz et al. 2018 .  ...  CSL dataset contains a smaller vocabulary compared with PHOENIX-2014. Following the works of (Huang et al. 2018; , the dataset is split by two strategies in Table 5 .  ... 
arXiv:2002.03187v1 fatcat:s4vr3lslfnfgzopf5pguy4ip5q

Quantitative Survey of the State of the Art in Sign Language Recognition [article]

Oscar Koller
2020 arXiv   pre-print
Additionally, it covers a fine-grained analysis on over 25 studies that have compared their recognition approaches on RWTH-PHOENIX-Weather 2014, the standard benchmark task of the field.  ...  Surprisingly, RWTH-PHOENIX-Weather with a vocabulary of 1080 signs represents the only resource for large vocabulary continuous sign language recognition benchmarking world wide.  ...  RWTH-PHOENIX-Weather 2014 [Koller et al., 2015] has a vocabulary of 1080 and the CSL corpus [Huang et al., 2018b] covers 178 signs.  ... 
arXiv:2008.09918v2 fatcat:quxu7qwjxvhl5nrerr5gesaboq

Spatial-Temporal Multi-Cue Network for Continuous Sign Language Recognition

Hao Zhou, Wengang Zhou, Yun Zhou, Houqiang Li
2020 PROCEEDINGS OF THE THIRTIETH AAAI CONFERENCE ON ARTIFICIAL INTELLIGENCE AND THE TWENTY-EIGHTH INNOVATIVE APPLICATIONS OF ARTIFICIAL INTELLIGENCE CONFERENCE  
To validate the effectiveness, we perform experiments on three large-scale CSLR benchmarks: PHOENIX-2014, CSL and PHOENIX-2014-T.  ...  We evaluate our method on three datasets, including PHOENIX-2014 (Koller, Forster, and Ney 2015) , CSL (Huang et al. 2018; Guo et al. 2018 ) and PHOENIX-2014-T (Cihan Camgoz et al. 2018) .  ...  CSL dataset contains a smaller vocabulary compared with PHOENIX-2014. Following the works of (Huang et al. 2018; Guo et al. 2018) , the dataset is split by two strategies in Table 5 .  ... 
doi:10.1609/aaai.v34i07.7001 fatcat:5zr6z3goifao5d45egwibcavpu
« Previous Showing results 1 — 15 out of 4,699 results