【今日CV 計算機視覺論文速覽】Wed, 13 Mar 2019

今日CS.CV計算機視覺論文速覽
Wed, 13 Mar 2019
Totally 25 papers

Interesting:

📖自動醫學圖像分析，主要就x光乳腺癌檢測，胸片CT肺結合檢測，腦部頸部病變檢測等方面展開研究，並闡述瞭如何生成數據、利用弱監督標籤、結合boosting方法等。（歐文分校博士論文）
1 Introduction
2 Adversarial Deep Structured Nets for Mass Segmentation from Mammograms
3 Deep Multi-instance Networks with Sparse Label Assignment for Whole Mammogram Classification
4 DeepLung: Deep 3D Dual Path Nets for Automated Pulmonary Nodule Detection and Classification
5 DeepEM: Deep 3D ConvNets With EM For Weakly Supervised Pulmonary Nodule Detection
6 AnatomyNet: Deep Learning for Fast and Fully Automated Whole-volume Segmentation of Head and Neck Anatomy
7 Conclusion


軟件著做：SOFTWARE
AnatomyNet
https://github.com/wentaozhu/AnatomyNet-for-anatomical-segmentation
Deep learning for fast and fully automated whole-volume segmentation of head and neck
anatomy.
DeepEM
https://github.com/wentaozhu/DeepEM-for-Weakly-Supervised-Detection
Deep 3D ConvNets with EM for weakly supervised pulmonary nodule detection.
DeepLung
xiv
https://github.com/wentaozhu/DeepLung
Deep 3d dual path nets for automated pulmonary nodule detection and classification.
Adversarial DSN
https://github.com/wentaozhu/adversarial-deep-structural-networks
Adversarial Deep Structural Networks for Mammographic Mass Segmentation.
Deep MIL
https://github.com/wentaozhu/deep-mil-for-whole-mammogram-classification
Deep multi-instance networks with sparse label assignment for whole mammogram classifi-
cation.
Regularized Deep LSTM
http://www.escience.cn/system/file?fileId=87579
Co-Occurrence feature learning for skeleton based action recognition using regularized deep
LSTM networks.

---

📖深度學習架構的理解類似性和差別性，主要集中於比較模型間和的類似性，研究發現核類似性達到99.9%的模型表現卻不盡相同，而沒用太大類似性的架構卻有着相同的表現！
20個網絡的相關性分析，表格中的數據爲模型在相同視覺智能和參數相關性上的差別：

12種resnet的相關性分析，

---

📖從RGB圖像生成複雜的形態學網格,主要集中與複雜形態的重建，提出了基於骨架媒介的和此方法，利用骨架保持形態並減少計算複雜度，並逐階段糾正體積和重建和網格的精調。（from 華南理工）

骨架體K合成的流程以及mesh優化的結果：

dataset：ShapeNet-Skeleton dataset
relate: 3D Shape net:http://3dshapenets.cs.princeton.edu/
3D Shape Synthesis and Recognition

---

📖GOGGLES,基於數據編程,多源數據的弱監督（data programming，label function）和類似性（affinity）編碼實現自動數據生成。


code: https://github.com/chu-data-lab/GOGGLES/
數據集：http://www.vision.caltech.edu/visipedia/CUB-200-2011.html

---

📖CaP，Cascaded Projection, 端到端的神經網絡壓縮和加速工具，基於數據驅動的方法實現，在保持高精度和高通量的前提下極大減少內存消耗。經過低秩投影的方法將輸入輸出連續的濾波器投影到統一的低維空間中來實現壓縮，並經過最小化分類損失和特徵間的差距來優化投影過程，並經過bp和sgd在幾何約束下獲得代理矩陣來進行優化，解決了精度、大小和速度的問題。（from rit Rochester Institute of Technology）

壓縮優化過程：

---

📖非監督的視頻顯著性物體檢測，主要利用了流補全的技術。首先利用光流檢測備選區域獲得光流邊界，隨後經過光流和補全流之間的差別獲得殘差流，並以此爲線索獲得運動顯著性掩膜。只依賴於運動信息讓這種方法具備靈活性和普適性。（from Inria, Centre Rennes）

相關方法和指標：

相關數據集和基準：https://davischallenge.org/davis2016/code.html

---

📖快速深度圖生成,提出了一種更高效的網絡架構用於從雙目視覺生成視差圖。研究人員利用了半分辨率的輸入，減少了網絡計算量，而且使用了低維（視差）特徵向量來實現了圖像間的配準。(from Swarthmore College,華盛頓大學)

code:https://projects.ayanc.org/fdscs/

---

📖Parallel Medical Imaging,(from 中科院自動化所)


數據合成方法值得注意：

!!!注：自動生成ct圖像

---

📖自適應遷移學習綜述，（from 重慶大學）

幾種自適應結構

---

📖手部骨骼分割，（from 新加坡國立）

dataset: 2017 Pediatric Bone Age Prediction Challenge [23]

---

📖利用深度學習生成超像素，（from KU Leuven）


相關：Simple Linear Iterative Clustering (SLIC)

Daily Computer Vision Papers

[1] Title: Dense Classification and Implanting for Few-Shot Learning
Authors:Yann Lifchitz, Yannis Avrithis, Sylvaine Picard, Andrei Bursuc
[2] Title: Placental Flattening via Volumetric Parameterization
Authors:S. Mazdak Abulnaga, Esra Abaci Turk, Mikhail Bessmeltsev, P. Ellen Grant, Justin Solomon, Polina Golland
[3] Title: An End-to-End Network for Panoptic Segmentation
Authors:Huanyu Liu, Chao Peng, Changqian Yu, Jingbo Wang, Xu Liu, Gang Yu, Wei Jiang
[4] **Title: Cascaded Projection: End-to-End Network Compression and Acceleration
Authors:Breton Minnehan, Andreas Savakis
[5] Title: Discriminative Principal Component Analysis: A REVERSE THINKING
Authors:Hanli Qiao
[6] **Title: Fast Deep Stereo with 2D Convolutional Processing of Cost Signatures
Authors:Kyle Yee, Ayan Chakrabarti
[7] Title: Hierarchical Autoregressive Image Models with Auxiliary Decoders
Authors:Jeffrey De Fauw, Sander Dieleman, Karen Simonyan
[8] Title: Parallel Medical Imaging: A New Data-Knowledge-Driven Evolutionary Framework for Medical Image Analysis
Authors:Chao Gou, Tianyu Shen, Wenbo Zheng, Oliver Kwan, Fei-Yue Wang
[9] Title: Unsupervised motion saliency map estimation based on optical flow inpainting
Authors:L. Maczyta, P. Bouthemy, O. Le Meur
[10] Title: Image Classification base on PCA of Multi-view Deep Representation
Authors:Yaoqi Sun, Liang Li, Liang Zheng, Ji Hu, Yatong Jiang, Chenggang Yan
[11] Title: Semi-Supervised Self-Taught Deep Learning for Finger Bones Segmentation
Authors:Ziyuan Zhao, Xiaoman Zhang, Cen Chen, Wei Li, Songyou Peng, Jie Wang, Xulei Yang, Le Zhang, Zeng Zeng
[12] Title: Paradox in Deep Neural Networks: Similar yet Different while Different yet Similar
Authors:Arash Akbarinia, Karl R. Gegenfurtner
[13] Title: Occlusion-guided compact template learning for ensemble deep network-based pose-invariant face recognition
Authors:Yuhang Wu, Ioannis A. Kakadiaris
[14] Title: Deep Learning for Automated Medical Image Analysis
Authors:Wentao Zhu
[15] Title: A Skeleton-bridged Deep Learning Approach for Generating Meshes of Complex Topologies from Single RGB Images
Authors:Jiapeng Tang, Xiaoguang Han, Junyi Pan, Kui Jia, Xin Tong
[16] Title: Knowledge Adaptation for Efficient Semantic Segmentation
Authors:Tong He, Chunhua Shen, Zhi Tian, Dong Gong, Changming Sun, Youliang Yan
[17] Title: Transfer Adaptation Learning: A Decade Survey
Authors:Lei Zhang
[18] Title: Fast Registration for cross-source point clouds by using weak regional affinity and pixel-wise refinement
Authors:Xiaoshui Huang, Lixin Fan, Qiang Wu, Jian Zhang, Chun Yuan
[19] Title: Quality-Gated Convolutional LSTM for Enhancing compressed video
Authors:Ren Yang, Xiaoyan Sun, Mai Xu, Wenjun Zeng
[20] Title: Generating superpixels using deep image representations
Authors:Thomas Verelst, Matthew Blaschko, Maxim Berman
[21] Title: GOGGLES: Automatic Training Data Generation with Affinity Coding
Authors:Nilaksh Das, Sanya Chaba, Sakshi Gandhi, Duen Horng Chau, Xu Chu
[22] Title: A total variation based regularizer promoting piecewise-Lipschitz reconstructions
Authors:Martin Burger, Yury Korolev, Carola-Bibiane Schönlieb, Christiane Stollenwerk
[23] Title: Generating Compact Geometric Track-Maps for Train Positioning Applications
Authors:Hanno Winter, Stefan Luthardt, Volker Willert, Jürgen Adamy
[24] Title: Theory III: Dynamics and Generalization in Deep Networks
Authors:Andrzej Banburski, Qianli Liao, Brando Miranda, Lorenzo Rosasco, Bob Liang, Jack Hidary, Tomaso Poggio
[25] Title: Progressive Generative Adversarial Binary Networks for Music Generation
Authors:Manan Oza, Himanshu Vaghela, Kriti Srivastava

Papers from arxiv.org

更多精彩請移步主頁

---

pic from pixels.com