Technology review – Face recognition with poor quality images

PDF

Authors
  1. Dong Yang, X.
  2. Dosselmann, R.
  3. Moise, M.
Corporate Authors
Defence R&D Canada - Ottawa, Ottawa ONT (CAN);Regina Univ, Regina Sask (CAN) Department of Computer Science
Abstract
Since the event of September 11th, 2001, research efforts in face recognition have intensified significantly, with much progress being made in the last decade in the essential areas of face detection and recognition algorithms. However, the development of face recognition systems that are truly useful and reliable in different and practical applications has been slow, mainly due to the limitations of existing algorithms. New challenges are raised in intelligence operations in which face images are captured in varying and unpredictable conditions. The resulting images usually contain many defects caused by poor or changing lighting, bad pose or camera angles that cause face occlusion and geometry changes. Worse yet, at a long distance, face images are sometimes of low resolution and contain high levels of noise. Before designing new application systems to deal with such situations, it is necessary to review and evaluate current face detection and facial recognition technologies. The objective of this technical report then is to review relevant research on facial recognition involving low resolution and/or poor quality images. We review a large number of scientific publications that address some of the questions above, albeit at different depths. This review focuses on the major topics and representative examples to date, so as to provide the reader with an understanding of the primary issues and current status of this field. It is not intended to be an exhaustive survey nor an

Il y a un résumé en français ici.

Report Number
DRDC-OTTAWA-CR-2012-012 — Contractor Report
Date of publication
01 May 2012
Number of Pages
44
DSTKIM No
CA036800
CANDIS No
536460
Format(s):
Electronic Document(PDF)

Permanent link

Document 1 of 1

Date modified: