Scalable face image retrieval with identity-based quantization and multi-reference re-ranking

State-of-the-art image retrieval systems achieve scala-bility by using bag-of-words representation and textual re-trieval methods, but their performance degrades quickly in the face image domain, mainly because they 1) produce vi-sual words with low discriminative power for face images, and 2) ignore the special properties of the faces. The lead-ing features for face recognition can achieve good retrieval performance, but these features are not suitable for inverted indexing as they are high-dimensional and global, thus not scalable in either computational or storage cost. In this paper we aim to build a scalable face image re-trieval system. For this purpose, we develop a new scal-able face representation using both local and global fea-tures. In the indexing stage, we exploit special proper-ties of faces to design new component-based local features, which are subsequently quantized into visual words using a novel identity-based quantization scheme. We also use a very small hamming signature (40 bytes) to encode the dis-criminative global feature for each face. In the retrieval stage, candidate images are firstly retrieved from the in-verted index of visual words. We then use a new multi-reference distance to re-rank the candidate images using the hamming signature. On a one-millon face database, we show that our local features and global hamming signa-tures are complementary—the inverted index based on local features provides candidate images with good recall, while the multi-reference re-ranking with global hamming signa-ture leads to good precision. As a result, our system is not only scalable but also outperforms the linear scan retrieval system using the state-of-the-art face recognition feature in term of the quality. 1