Large-Scale Face Image Retrieval using Semantic Codewords
碩士 === 國立臺灣大學 === 資訊工程學研究所 === 100 === Photos with people (e.g., family, friends, celebrities, etc.) are the ma- jor interest of users. Thus, with the exponentially growing photos, large- scale content-based face image retrieval is an enabling technology for many emerging applications. In this work,...
| Main Authors: | , |
|---|---|
| Other Authors: | |
| Format: | Others |
| Language: | en_US |
| Published: |
2012
|
| Online Access: | http://ndltd.ncl.edu.tw/handle/56924292502137142665 |
| Summary: | 碩士 === 國立臺灣大學 === 資訊工程學研究所 === 100 === Photos with people (e.g., family, friends, celebrities, etc.) are the ma- jor interest of users. Thus, with the exponentially growing photos, large- scale content-based face image retrieval is an enabling technology for many emerging applications. In this work, we aim to develop a scalable face image retrieval system which can integrate with auxiliary information to improve the retrieval result. To achieve this goal, we first apply sparse coding on local features extracted from face images combining with inverted indexing to construct an efficient and scalable face retrieval system. We then propose two different coding scheme that utilize partial identity information and automatically detected human attributes to construct semantic codewords for further improving the retrieval results. Using the proposed coding schemes, face images with large intra-class variances will still be quantized into similar semantic codewords if they share the same identity or similar human attributes. We investigate the effectiveness of different attributes and vital factors essen- tial for face retrieval. Experimental results show that the proposed methods can achieve salient retrieval results compared to existing methods in two public datasets.
|
|---|