【今日CS 視覺論文速覽】3 Jan 2019
今日CS.CV計算機視覺論文速覽
Thu, 3 Jan 2019
Totally 38 papers
Interesting:
-
將古代花鳥山水轉換為照片的風格遷移,通過域遷移的方法將古畫處理問題轉變成了自然影象處理問題,在自然影象上訓練的模型可以應用到遷移繪畫中,在古畫中對真實照片訓練的分類模型和風格模型進行了遷移。研究人員主要收集了宋代、清代的花鳥和山水畫資料集,並建立了域風格遷移網路。通過複雜的損失函式保證了遷移後的影象保持源影象的色彩和內容。(from 浙江大學)
研究人員收集的三個資料集,其中古畫圖片花(2258+650)鳥(2119+600)山水(2009+600):
採用的網路結構:
最終得到的結果:
dataset:CFP,CBP,CLP; 花朵分類器:Oxford Flower;語義分割任務:PASCAL VOC
2012
ref:https://person.zju.edu.cn/0092050 -
EdgeConnect,一種基於邊緣補全的影象修復新方法,這篇文章將影象修復的工作分成了兩個部分,首先利用利用啟發式的生成模型得到了缺失部分的邊緣資訊,隨後將邊緣資訊作為影象缺失的先驗部分和影象一起送入修復網路進行影象重建。(from 安大略技術大學)
感受一下效果:
dataset:CelebA, Places2, and Paris Street View
Code:https://github.com/knazeri/edge-connect
related inpainting:
https://github.com/satoshiiizuka/siggraph2017_inpainting
https://github.com/JiahuiYu/generative_inpainting -
掩膜輔助的人群計數方法,由於人群估計的問題主要在於密度估計,而在掩膜的加入可以減小密度估計的難度,同時掩膜估計問題又可以轉換為二值化的分割問題來解決。在傳統方法的基礎上增加了目標掩膜的分支,隨後將預測出的掩膜與與輸入圖結合生成更好的密度圖。(from 南京大學 阿德萊德大學 澳大利亞)
研究人員提出了五種不同的架構來實現mask的預測和融合預測密度圖的方式:
人群計數資料集: shanghaitech, UCF_CC_50, WorldExpo10, The MALL
ref:http://cs-chan.com/downloads_crowd_dataset.html
https://github.com/svishwa/crowdcount-mcnn
https://irc.atr.jp/sets/TEMPOSAN_dataset/
港中文的大資料集 -
Action2Vec,建立了銜接語言資訊和視覺空間資訊的嵌入隱含空間,將動作和語言描述用類似word2vec的方式銜接起來。(from 佐治亞理工)
嵌入空間的視覺化:
同時在嵌入空間中實現了代數運算,對動作和主體進行了代數操作:
dataset:UCF101 [29], HMDB51[18] and Kinetics [13]. -
學習三維剛體的物理動力學過程,通過輸入目標點雲、衝量向量得到了物體在物理環境中受力作用後的最終位姿,這一模型的物理動力學學習結果還能用於未知物體的動力學估計。(from 斯坦福)
網路模型,輸入物體點雲和輸出的力通過綜合後得到物體的最終位姿:
dataset:ShapeNet
模擬環境:
https://pybullet.org/
https://unity3d.com/
Author:
https://github.com/davrempe
https://cs.stanford.edu/people/ssrinath/
https://geometry.stanford.edu/member/guibas/index.html
The hierarchical relation network -
利用模糊資料來訓練模型,保護使用者隱私,利用人眼難以分辨但是機器可以使用的影象來訓練演算法。在分類、屬性分類和人臉關鍵點檢測方面取得了不錯的結果。通過訓練模糊網路來處理資料,隨後利用處理的資料來訓練目標網路。
-
(from Deeping Source)
![![在這裡插入圖片描述](https://img-blog.csdnimg.cn/20190104174700976. =500x)
檢測資料集:SVHN, CIFAR10, Pascal VOC 2012, CelebA, and MTFL.
ref:http://www.deepingsource.io/ -
SiCloPe,單張影象生成人體衣著旋轉效果的模型,基於模特的剪影研究人員可以通過這一模型重建人體衣著的三維模型。這意味著在虛擬試裝時可以看到自己前後左右的衣著效果。這一工作利用了二維剪影和三維關節位置資料來描述複雜變化的人體穿著場景。首先通過利用輸入剪影和關節資料合成了新視角下連續的剪影,隨後利用生成網路得到目標的三維模型。最後利用前檢視生成後檢視,從而得到紋理來對三維模型的表面進行處理。(from 美國南加州大學創意技術研究所)
新視角下的剪影合成網路:
前後對映模型:
一些結果:
dataset:rigged meshes,aXYZ, Renderpeople, animation sequences Mixamo, HDRI Haven -
SIXray,提出了一個大規模的安檢X光資料集,包含了1059231張X光安檢資料,並對其中的6類共8929個違禁品進行了手動標記。其特點是很多物體之間有遮擋關係。研究人員提出了類平衡的層級精煉方法來處理複雜物件和資料不平衡的情況,同時引入了高階視覺特徵輔助中級特徵。利用中特徵檢測得到了很好地效果,使得弱監督學習成為可能。(from 中科大)
資料集由不同層的透明影象疊加構成:
論文中提出的層級平衡精煉方法:
一些檢測到違禁品的結果:
安檢X光資料集SIXray,ref:GDXray -
一種字元檢測的方法,(from百度)
文字字元Text檢測資料集:The VGG SynthText dataset, ICDAR13, MSRA-TD500.,Total-Text
文字字元識別比賽會議ref:http://u-pat.org/ICDAR2017/index.php
http://u-pat.org/ICDAR2017/program_competitions.php
http://u-pat.org/ICDAR2017/index.php
http://rrc.cvc.uab.es/
http://tc11.cvc.uab.es/datasets/icdar15smartdoc-ch2_1
https://arxiv.org/pdf/1601.07140.pdf -
利用3D合成法生成人臉欺詐資料集,利用列印的彩色頭像轉換為三維網格,並進行隨機的彎曲和選擇,最後利用透視變換渲染出虛擬的樣本。(from 中科大)
-
多輸出學習的綜述,(from 悉尼技術大學)
-
基於FPGA加速的深度學習綜述,(from 法赫德國王石油礦產大學,沙特)
-
Lipi Gnani,一個印度卡納達語的字元識別轉換系統,(from 印度科學院)
Daily Computer Vision Papers
[1] Title: Improving Face Anti-Spoofing by 3D Virtual Synthesis
Authors:Jianzhu Guo, Xiangyu Zhu, Jinchuan Xiao, Zhen Lei, Genxun Wan, Stan Z. Li
[2] Title: Action2Vec: A Crossmodal Embedding Approach to Action Learning
Authors:Meera Hahn, Andrew Silva, James M. Rehg
[3] Title: Learning Generalizable Physical Dynamics of 3D Rigid Objects
Authors:Davis Rempe, Srinath Sridhar, He Wang, Leonidas J. Guibas
[4] Title: Improved Hyperspectral Unmixing With Endmember Variability Parametrized Using an Interpolated Scaling Tensor
Authors:Ricardo Augusto Borsoi, Tales Imbiriba, José Carlos Moreira Bermudez
[5] Title: Lipi Gnani - A Versatile OCR for Documents in any Language Printed in Kannada Script
Authors:Shiva Kumar H R, Ramakrishnan A G
[6] Title: Attribute-Aware Attention Model for Fine-grained Representation Learning
Authors:Kai Han, Jianyuan Guo, Chao Zhang, Mingjian Zhu
[7] Title: Learning Efficient Detector with Semi-supervised Adaptive Distillation
Authors:Shitao Tang, Litong Feng, Wenqi Shao, Zhanghui Kuang, Wei Zhang, Yimin Chen
[8] Title: Detecting Text in the Wild with Deep Character Embedding Network
Authors:Jiaming Liu, Chengquan Zhang, Yipeng Sun, Junyu Han, Errui Ding
[9] Title: Optical Fringe Patterns Filtering Based on Multi-Stage Convolution Neural Network
Authors:Bowen Lin, Shujun Fu, Caiming Zhang, Fengling Wang, Yuliang Li
[10] Title: Plugin Networks for Inference under Partial Evidence
Authors:Michal Koperski, Tomasz Konopczynski, Piotr Semberecki, Tomasz Trzcinski
[11] Title: SIXray : A Large-scale Security Inspection X-ray Benchmark for Prohibited Item Discovery in Overlapping Images
Authors:Caijing Miao, Lingxi Xie, Fang Wan, Chi Su, Hongye Liu, Jianbin Jiao, Qixiang Ye
[12] Title: On Minimum Discrepancy Estimation for Deep Domain Adaptation
Authors:Mohammad Mahfujur Rahman, Clinton Fookes, Mahsa Baktashmotlagh, Sridha Sridharan
[13] Title: Vector and Line Quantization for Billion-scale Similarity Search on GPUs
Authors:Wei Chen, Jincai Chen, Fuhao Zou, Yuan-Fang Li, Ping Lu, Qiang Wang, Wei Zhao
[14] Title: Ancient Painting to Natural Image: A New Solution for Painting Processing
Authors:Tingting Qiao, Weijing Zhang, Miao Zhang, Zixuan Ma, Duanqing Xu
[15] Title: EdgeConnect: Generative Image Inpainting with Adversarial Edge Learning
Authors:Kamyar Nazeri, Eric Ng, Tony Joseph, Faisal Qureshi, Mehran Ebrahimi
[16] Title: Mapping Areas using Computer Vision Algorithms and Drones
Authors:Bashar Alhafni, Saulo Fernando Guedes, Lays Cavalcante Ribeiro, Juhyun Park, Jeongkyu Lee
[17] Title: Nasal Patches and Curves for Expression-robust 3D Face Recognition
Authors:Mehryar Emambakhsh, Adrian Evans
[18] Title: Handwritten Indic Character Recognition using Capsule Networks
Authors:Bodhisatwa Mandal, Suvam Dubey, Swarnendu Ghosh, Ritesh Sarkhel, Nibaran Das
[19] Title: Rethinking on Multi-Stage Networks for Human Pose Estimation
Authors:Wenbo Li, Zhicheng Wang, Binyi Yin, Qixiang Peng, Yuming Du, Tianzi Xiao, Gang Yu, Hongtao Lu, Yichen Wei, Jian Su
[20] Title: Gated-Dilated Networks for Lung Nodule Classification in CT scans
Authors:Mundher Al-Shabi, Hwee Kuan Lee, Maxine Tan
[21] Title: Training with the Invisibles: Obfuscating Images to Share Safely for Learning Visual Recognition Models
Authors:Tae-hoon Kim, Dongmin Kang, Kari Pulli, Jonghyun Choi
[22] Title: Not All Words are Equal: Video-specific Information Loss for Video Captioning
Authors:Jiarong Dong, Ke Gao, Xiaokai Chen, Junbo Guo, Juan Cao, Yongdong Zhang
[23] Title: Extreme Relative Pose Estimation for RGB-D Scans via Scene Completion
Authors:Zhenpei Yang, Jeffrey Z.Pan, Linjie Luo, Xiaowei Zhou, Kristen Grauman, Qixing Huang
[24] Title: Multiple Sclerosis Lesion Inpainting Using Non-Local Partial Convolutions
Authors:Hao Xiong, Dacheng Tao
[25] Title: A Noise-Sensitivity-Analysis-Based Test Prioritization Technique for Deep Neural Networks
Authors:Long Zhang, Xuechao Sun, Yong Li, Zhenyu Zhang, Yang Feng
[26] Title: SiCloPe: Silhouette-Based Clothed People
Authors:Ryota Natsume, Shunsuke Saito, Zeng Huang, Weikai Chen, Chongyang Ma, Hao Li, Shigeo Morishima
[27] Title: Deep Information Theoretic Registration
Authors:Alireza Sedghi, Jie Luo, Alireza Mehrtash, Steve Pieper, Clare M. Tempany, Tina Kapur, Parvin Mousavi, William M. Wells III
[28] Title: Mask-aware networks for crowd counting
Authors:Shengqin Jiang, Xiaobo Lu, Yinjie Lei, Lingqiao Liu
[29] Title: Interest Point Detection based on Adaptive Ternary Coding
Authors:Zhenwei Miao, Kim-Hui Yap, Xudong Jiang
[30] Title: DCI: Discriminative and Contrast Invertible Descriptor
Authors:Zhenwei Miao, Kim-Hui Yap, Xudong Jiang, Subbhuraam Sinduja, Zhenhua Wang
[31] Title: Learning Spatial Common Sense with Geometry-Aware Recurrent Networks
Authors:Hsiao-Yu Fish Tung, Ricson Cheng, Katerina Fragkiadaki
[32] Title: Impact of Ground Truth Annotation Quality on Performance of Semantic Image Segmentation of Traffic Conditions
Authors:Vlad Taran, Yuri Gordienko, Alexandr Rokovyi, Oleg Alienin, Sergii Stirenko
[33] Title: Instant Automated Inference of Perceived Mental Stress through Smartphone PPG and Thermal Imaging
Authors:Youngjun Cho, Simon J. Julier, Nadia Bianchi-Berthouze
[34] Title: AVRA: Automatic Visual Ratings of Atrophy from MRI images using Recurrent Convolutional Neural Networks
Authors:Gustav Mårtensson, Daniel Ferreira, Lena Cavallin, J-Sebastian Muehlboeck, Lars-Olof Wahlund, Chunliang Wang, Eric Westman
[35] Title: A Survey on Multi-output Learning
Authors:Donna Xu, Yaxin Shi, Ivor W. Tsang, Yew-Soon Ong, Chen Gong, Xiaobo Shen
[36] Title: FPGA-based Accelerators of Deep Learning Networks for Learning and Classification: A Review
Authors:Ahmad Shawahna, Sadiq M. Sait, Aiman El-Maleh
[37] Title: Dense Morphological Network: An Universal Function Approximator
Authors:Ranjan Mondal, Sanchayan Santra, Bhabatosh Chanda
[38] Title: Deep Frame Prediction for Video Coding
Authors:Hyomin Choi, Ivan V. Bajic