【今日CS 視覺論文速覽】4 Jan 2019
今日CS.CV計算機視覺論文速覽
Fri, 4 Jan 2019
Totally 20 papers
Interesting:
-
雜貨店資料集,包含了層級歸類資訊、視覺資訊和語義標籤。包含了自然場景和獨立的影象,共有81個細粒度的分類和5125張影象,每類包含30-138張。(from 瑞典皇家理工 微軟劍橋)
粗分類和細分類:
用於評測的分類方法:
同時還進行了從自然影象到獨立影象間的編碼解碼生成:
資料集地址:https://github.com/marcusklasson/GroceryStoreDataset
ref:Flickr30k FIDS30 -
GeoNet,第一個基於點雲來為物體表面固有(Geodesic)結構建模的神經網路。研究人員提出了一種融合機制利用GeoNet與基礎網路結合起來對點雲進行分析。這篇文章提出了GeoNet來從點雲資料中學習出表面拓撲形態的表示方法,並用測地線(geodesic)融合的方法對學習到的形態學特徵進行分析和利用。(from UCLA Face++ 斯坦福 UIUC)
其中GeoNet,測地線鄰域估計網路的結果如下圖所示,基於自編碼器抽取每點特徵並利用測地線匹配來作為核函式,通過學習到的嵌入特徵來估計測地線的鄰域資訊。GM層的中間特徵包含了豐富的點雲形態和表面資訊。:
下圖是測地線融合機制,可用於點雲加密上取樣、法向量重建和非剛體的形狀分類等。
與PU-Net融合後點雲上取樣的結果:
與PointNET++融合後法向量估計和泊松表面重建結果。
長寬比過大的條狀物失敗樣本:
三維點雲資料集:ShapeNet, Shrec15
ref:
15dataset:http://www.cs.cf.ac.uk/shaperetrieval/shrec15/SHREC15.zip
三維形狀檢索:http://www.shrec.net/
Office web:http://www.shrec.net/
三維草圖檢索:https://userweb.cs.txstate.edu/~yl12/SBR2016/
2015:http://www.cs.cf.ac.uk/shaperetrieval/shrec15/
2018:http://www.shrec.net/
歐洲圖形學會:https://www.eg.org/wp/
一作:https://tonghehehe.com/
二作:https://people.cs.umass.edu/~hbhuang/(三維點雲方向) -
單位球空間中的體卷積,將歐式空間中的卷積擴充套件到了單位球空間中。(from 澳大利亞國立大學)
球面和求空間卷積比較:
一個使用三維球卷積做分類的例子,p+p-代表水平y方向正負兩個點雲:
在ModelNet10和ModelNet40上和一些三維方法的比較,層數少效率優勢明顯:
資料集:ModelNet
uniball:https://en.wikipedia.org/wiki/Uni-ball
Zernike_polynomials:https://en.wikipedia.org/wiki/Zernike_polynomials -
一個用於衛星影象去雲朵遮擋的資料集Remote sensing Image Cloud rEmoving dataset (RICE),其中RICE1包含了500對512*512大小有云和無雲的資料,RICE2則包含了450個,每個由三張影象(雲圖,雲mask圖和無雲的圖)組成的系列影象。
帶label和mask的RICE2 資料:
dataset:https://github.com/BUPTLdy/RICE_DATASET -
基於光流的自監督影象分割方法, 利用合成光流資料訓練光流估計器,隨後利用光流估計器在真實資料中學習出影象的特徵,隨後將這些特徵應用於影象分割任務。文章最大的貢獻在於利用合成數據估計以及通過運動線索motion cue估計影象特徵。(from CuraCloud)
在人體血管分割上的效果:
ref:PWC-Net
資料集:X-ray coronary angiography -
線性顏色分割問題,基於局域鄰域圖框架,並提出了邊緣權重函式。(from 俄羅斯科學院)
ref:
code:http://github.com/visillect/segmentation
dataset:Barnard’s DXC-930 SFU dataset
資料集:CLEVR-Ref+: Diagnosing Visual Reasoning with Referring Expressions
https://www.jianshu.com/p/321f39b7fa93 -
人臉識別綜述,基於多級分類法來進行綜述。基於各種人臉特徵總結了目前的方法,並提出了一系列挑戰。(from里斯本大學)
人臉資料集:
Daily Computer Vision Papers
[1] Title: CLEVR-Ref+: Diagnosing Visual Reasoning with Referring Expressions
Authors:Runtao Liu, Chenxi Liu, Yutong Bai, Alan Yuille
[2] Title: Face Recognition: A Novel Multi-Level Taxonomy based Survey
Authors:Alireza Sepas-Moghaddam, Fernando Pereira, Paulo Lobato Correia
[3] Title: A Hierarchical Grocery Store Image Dataset with Visual and Semantic Labels
Authors:Marcus Klasson, Cheng Zhang, Hedvig Kjellström
[4] Title: Generating Multiple Objects at Spatially Distinct Locations
Authors:Tobias Hinz, Stefan Heinrich, Stefan Wermter
[5] Title: GeoNet: Deep Geodesic Networks for Point Cloud Analysis
Authors:Tong He, Haibin Huang, Li Yi, Yuqian Zhou, Chihao Wu, Jue Wang, Stefano Soatto
[6] Title: Active Learning with TensorBoard Projector
Authors:Francois Luus, Naweed Khan, Ismail Akhalwaya
[7] Title: Baseline Desensitizing In Translation Averaging
Authors:Bingbing Zhuang, Loong-Fah Cheong, Gim Hee Lee
[8] Title: Edge-Semantic Learning Strategy for Layout Estimation in Indoor Environment
Authors:Weidong Zhang, Wei Zhang, Jason Gu
[9] Title: A Remote Sensing Image Dataset for Cloud Removal
Authors:Daoyu Lin, Guangluan Xu, Xiaoke Wang, Yang Wang, Xian Sun, Kun Fu
[10] Title: Adaptive Locality Preserving Regression
Authors:Jie Wen, Zuofeng Zhong, Zheng Zhang, Lunke Fei, Zhihui Lai, Runze Chen
[11] Title: Photo-Sketching: Inferring Contour Drawings from Images
Authors:Mengtian Li, Zhe Lin, Radomir Mech, Ersin Yumer, Deva Ramanan
[12] Title: Visualizing Deep Similarity Networks
Authors:Abby Stylianou, Richard Souvenir, Robert Pless
[13] Title: Linear colour segmentation revisited
Authors:Anna Smagina, Valentina Bozhkova, Sergey Gladilin, Dmitry Nikolaev
[14] Title: Flow Based Self-supervised Pixel Embedding for Image Segmentation
Authors:Bin Ma, Shubao Liu, Yingxuan Zhi, Qi Song
[15] Title: Low-Cost Device Prototype for Automatic Medical Diagnosis Using Deep Learning Methods
Authors:Neil Deshmukh
[16] Title: Resource-Scalable CNN Synthesis for IoT Applications
Authors:Mohammad Motamedi, Felix Portillo, Mahya Saffarpour, Daniel Fong, Soheil Ghiasi
[17] Title: Topology, homogeneity and scale factors for object detection: application of eCognition software for urban mapping using multispectral satellite image
Authors:Polina Lemenkova
[18] Title: Volumetric Convolution: Automatic Representation Learning in Unit Ball
Authors:Sameera Ramasinghe, Salman Khan, Nick Barnes
[19] Title: Multi-class Classification without Multi-class Labels
Authors:Yen-Chang Hsu, Zhaoyang Lv, Joel Schlosser, Phillip Odom, Zsolt Kira
[20] Title: Honey Authentication with Machine Learning Augmented Bright-Field Microscopy
Authors:Peter He, Alexis Gkantiragas, Gerard Glowacki