今日から始める人工知能 × 機械学習 Meetup ライトニングトーク1

今日から始める人工知能 × 機械学習 Meetup ライトニングトーク1

• 1.
  Large-scale Semantic VisualSearch NGUYEN ANH TUAN [email protected] 2016/07/17
• 2.
  About me • 東京大学情報理工学系研究科 修士2年生 • テーマ：Object Retrieval，情 報検索等 • 趣味：水泳，囲碁 • ブログ： https://imsmarxen68.tumblr.co m/ NGUYEN ANH TUAN 東京大学・情報理 工・修士2年生
• 3.
  A picture isworth a thousand words
• 4.
  Outline • Semantic VisualSearch • A visual search framework Image credits: http://ai.stanford.edu/~jkrause/cars/car_dataset.html Feature extraction Feature aggregation Feature matching Re-ranking Preliminary results Final results NGUYEN ANH TUAN 東京大学・情報理 工・修士2年生
• 5.
  Visual search Image credits:http://ai.stanford.edu/~jkrause/cars/car_dataset.html Image credits: http://google.com NGUYEN ANH TUAN 東京大学・情報理 工・修士2年生
• 6.
  What’s the problem? •Semantic difficulties: fine-grained differences Image credits: http://ai.stanford.edu/~jkrause/cars/car_dataset.html NGUYEN ANH TUAN 東京大学・情報理 工・修士2年生
• 7.
  But for searchproblem? Image credits: http://ai.stanford.edu/~jkrause/cars/car_dataset.html Query Database NGUYEN ANH TUAN 東京大学・情報理 工・修士2年生
• 8.
  But for searchproblem? Image credits: http://ai.stanford.edu/~jkrause/cars/car_dataset.html Query Database 0.1 0.5 0.2Ranking problem with a variation of fine-grained changes NGUYEN ANH TUAN 東京大学・情報理 工・修士2年生
• 9.
  But for searchproblem? Image credits: http://ai.stanford.edu/~jkrause/cars/car_dataset.html Query Database 0.1 0.5 0.2Find visual representations to capture all fine-grained local information in images NGUYEN ANH TUAN 東京大学・情報理 工・修士2年生
• 10.
  Large-scale Visual Search
• 11.
  Robust feature extraction •Robust to – Scale changes – Rotation and affine changes – Blur, sharpening, … Feature extraction Feature aggregation Feature matching Re-ranking Preliminary results Final results Image credits: http://ai.stanford.edu/~jkrause/cars/car_dataset.html A picture is worth a thousand words NGUYEN ANH TUAN 東京大学・情報理 工・修士2年生
• 12.
  Statistical kernels • Bag-of-Features(BoF) • Fisher kernel (GMM) [1] • VLAD (K-means) [2] Image credits: http://www.mathworks.com/matlabcentral/ Feature extraction Feature aggregation Feature matching Re-ranking Preliminary results Final results [1] F. Perronnin, C. Dance, “Fisher Kernels on Visual Vocabularies for Image Categorization,” in Proc. CVPR, IEEE, 2007 [2] H. Jegou, F. Perronnin, M. Douze, J. Sanchez, P. Perez, C. Schmid, “Aggregating Local Image Descriptors into Compact Codes,” IEEE Trans. Pattern Anal. Mach. Intell. 34 (2012) 1704–1716. NGUYEN ANH TUAN 東京大学・情報理 工・修士2年生
• 13.
  Statistical kernels Feature extraction Feature aggregation Feature matching Re-ranking Preliminary results Final results NGUYENANH TUAN 東京大学・情報理 工・修士2年生
• 14.
  Image matching =Feature matching • Feature matching→Nearest Neighbor Search – Inverse Search with Inverted Indices – Compressed data for better memory usage [3] Feature extraction Feature aggregation Feature matching Re-ranking Preliminary results Final results [3] H. Jégou, M. Douze, C. Schmid, Product quantization for nearest neighbor search., IEEE Trans. Pattern Anal. Mach. Intell. 33 (2011) 117– 28.Data CompressionNGUYEN ANH TUAN 東京大学・情報理 工・修士2年生
• 15.
  Verification • Geometry verification –RANSAC methods [4] – Reduce the number of good inliers Image credits: http://ai.stanford.edu/~jkrause/cars/car_dataset.html Feature extraction Feature aggregation Feature matching Re-ranking Preliminary results Final results [4] M.A. Fischler, R.C. Bolles, Random sample consensus: a paradigm for model fitting with applications to image analysis and automated cartography, Commun. ACM. 24 (1981) 381–395. NGUYEN ANH TUAN 東京大学・情報理 工・修士2年生
• 16.
  Thank you forlistening