[2018.05.09] paper skim

[unizard]

정리요망

Private Interest

• FFNet: Video Fast-Forwarding via Reinforcement Learning
• A Memory Network Approach for Story-based Temporal Summarization of 360◦ Videos
  • Memory Network, 360 Videos, SNU
• Learning to See in the Dark
  • Fun, INtel zz

Focussing Technical Method

• Advanced local motion patterns for macro and micro facial expression recognition, !how to extract macro & micro movement

• A Hierarchical Matcher using Local Classifier Chains

  • hierarchical matcher

• Deep Ordinal Hashing with Spatial Attention

  • spatial attention

• Visual Global Localization with a Hybrid WNN-CNN Approach

  • How to describe Global & Local

• Learning image-to-image translation using paired and unpaired training samples

  • Paired and Unpaired

• Weakly-supervised Visual Instrument-playing Action Detection in Videos

  • DB, Code, PlayAction Detection

• Weakly-Supervised Video Object Grounding from Text by Loss Weighting and Object Interaction

  • 텍스트의 대상이 되는 비디오 물체 CAM, 민수님 연구와 비슷

• Position Estimation of Camera Based on Unsupervised Learning

  • Camera Parameter

• Reliability Map Estimation For CNN-Based Camera Model Attribution

  • Camera Parameter

• Robust Classification with Convolutional Prototype Learning

  • Class Incremental Learning ??

시간이없어서정리를못하네 #흑흑 #시간아멈추어다오