「ECCV 2016」Reading Notes
一、檢測、識別和檢索（Detection, Recognition and Retrival）
SSD: Single Shot MultiBox Detector ECCV 2016
P1, pp. 21-37
二、Poster Session 1
Learning to Refine Object Segments
P1, pp. 75-91
Deep Automatic Portrait Matting
P1, pp. 92-107
Region-Based Semantic Segmentation with End-to-End Training
P1, pp. 381-397
Convolutional Oriented Boundaries
P1, pp. 580-596
Superpixel Convolutional Networks Using Bilateral Inceptions
P1, pp. 597-613
三、場景理解（Scene Understanding）
四、Poster Session 2
Single Image Dehazing via Multi-scale Convolutional Neural Networks
P2, pp. 154-169
A Cluster Sampling Method for Image Matting via Sparse Coding
P2, pp. 204-219
Accelerating the Super-Resolution Convolutional Neural Network
P2, pp. 391-407
Higher Order Conditional Random Fields in Deep Neural Networks
P2, pp. 524-540
Natural Image Matting Using Deep Convolutional Neural Networks
P2, pp. 626-643
Perceptual Losses for Real-Time Style Transfer and Super-Resolution
P2, pp. 694-711
五、Optimization（優化）
六、Poster Session 3
A Neural Approach to Blind Motion Deblurring
P3, pp. 221-235
Laplacian Pyramid Reconstruction and Refinement for Semantic Segmentation
P3, pp 519-534
七、影象和視訊處理（Image and Video Processing）
The Fast Bilateral Solver
P3, pp. 617-632
Colorful Image Colorization
P3, pp. 649-666
八、Poster Session 4
Precomputed Real-Time Texture Synthesis with Markovian Generative Adversarial Networks
P3, pp. 702-716
HFS: Hierarchical Feature Selection for Efficient Image Segmentation
P3, pp. 867-882
Deep Joint Image Filtering
P4, pp. 154-169
Generative Image Modeling Using Style and Structure Adversarial Networks
P4, pp. 318-335
A Unified Multi-scale Deep Convolutional Neural Network for Fast Object Detection
P4, pp. 354-370
九、學習（Learning）
Learning Representations for Automatic Colorization
P4, pp. 577-593
Identity Mappings in Deep Residual Networks
P4, pp. 630-645
Deep Networks with Stochastic Depth
P4, pp. 646-661