survey.md

Face Recognition Survey

• FaceNet: A Unified Embedding for Face Recognition and Clustering
  • CVPR 2015
  • Triplet Lossを用いた128次元の顔画像embeddingの学習
  • 学習中、オンラインでTriplet(Hard NegativeとHard Positive)を選択
  • Source1
  • Source2
  • Source3
• MobileID: Face Model Compression by Distilling Knowledge from Neurons
  • AAAI 2016
  • Source
• Triplet Probabilistic Embedding for Face Verification and Clustering
  • BTAS 2016
  • Source
• A Discriminative Feature Learning Approach for Deep Face Recognition
  • ECCV 2016
  • Softmax Loss + Center Loss
  • Source
  • Source
• Pose-Robust Face Recognition via Deep Residual Equivariant Mapping
  • CVPR 2018
  • 横顔を正面顔にMappingするDREAM Blockの導入により、顔の姿勢にロバストな顔認識
  • Source
• MobileFaceNets: Efficient CNNs for Accurate RealTime Face Verification on Mobile Devices
  • CCBR 2018
  • MobileNetV2を元にした顔画像embedding学習
  • 携帯電話向けの早いモデル
  • Source
• ArcFace: Additive Angular Margin Loss for Deep Face Recognition
  • Source
• CosFace: Large Margin Cosine Loss for Deep Face Recognition
  • CVPR 2018
  • LossをCosine距離で計算する際、クラス間にmarginをつける
  • OSSにはLicense付けられてない
  • Source

Hashing high-dimensional vector

• FAst Lookups of Cosine and Other Nearest Neighbors

  • Source

• [Locality Sensitive Hashing using MinHash]

  • Source

• [Approximate Nearest Neighbors in C++/Python optimized for memory usage and loading/saving to disk]

  • Spotifyが開発した音楽推薦システムのためのApproximate Nearest Neighbors
  • Sources