Construction of a large scale image feature database using sparse coding
Project/Area Number |
25730070
|
Research Category |
Grant-in-Aid for Young Scientists (B)
|
Allocation Type | Multi-year Fund |
Research Field |
Multimedia database
|
Research Institution | University of Tsukuba |
Principal Investigator |
TEZUKA Taro 筑波大学, 図書館情報メディア系, 准教授 (40423016)
|
Project Period (FY) |
2013-04-01 – 2016-03-31
|
Project Status |
Completed (Fiscal Year 2016)
|
Budget Amount *help |
¥4,160,000 (Direct Cost: ¥3,200,000、Indirect Cost: ¥960,000)
Fiscal Year 2016: ¥780,000 (Direct Cost: ¥600,000、Indirect Cost: ¥180,000)
Fiscal Year 2015: ¥1,040,000 (Direct Cost: ¥800,000、Indirect Cost: ¥240,000)
Fiscal Year 2014: ¥1,170,000 (Direct Cost: ¥900,000、Indirect Cost: ¥270,000)
Fiscal Year 2013: ¥1,170,000 (Direct Cost: ¥900,000、Indirect Cost: ¥270,000)
|
Keywords | スパース符号化 / 辞書学習 / 画像特徴量 / 画像認識 / スパース性 / 画像分類 / 画像処理 |
Outline of Final Research Achievements |
Dictionary learning is a general name given to methods that features for decomposing observed data into a sparse linear combination. One achievement of this project was to develop a method of speeding up K-SVD, a widely-used method for dictionary learning, by iterative projection onto lower dimensional subspace. The result was published as an article in the Journal of Information Processing by the Information Processing Society of Japan in May 2016. By applying the method to a set of images, an image feature database was constructed. In addition, a method of image classification using a Fisher kernel imposing sparsity was developed.
|
Report
(4 results)
Research Products
(19 results)