Face-based Document Image Retrieval System

Abstract

Many of the documents such as passport, identity card, property documents, voter ID and certificates include face image of a person. These types of documents are used in various off-line and on-line applications. Retrieval of such documents based on face image is a useful application in government offices, various companies, and organizations. This paper proposes a method to retrieve documents from the database pertaining to a person based on face image. The proposed method includes three steps: (1) Detecting face image from query document (ii) Extracting the face image features and (iii) Retrieving all documents that contain face image similar to that of the query document. The proposed method extends the concept of grey level co-occurrence matrix to color images for feature extraction. Experiments are carried out on the database of 810 document images, in which face images are borrowed from face94 [1]. Mean Average Precision (MAP) of 82.66% is achieved with proposed method and the results are compared with current state of art including Haralick features associated with Grey Level Co-occurrence Matrix (GLCM) [2]. It is found that the results are encouraging compared with existing methods.