Temporal Motion Vector Filter for Fast Object Detection on Compressed Video

  • Ronaldo Carvalho Moura Monity
  • Elder Moreira Hemerly Instituto Tecnológico de Aeronáutica
  • Adilson Marques da Cunha Instituto Tecnológico de Aeronáutica

Abstract

A novel Temporal Motion Vector Filter (TF) is presented and evaluated for real-time object detection on compressed videos in MPEG-2, MPEG-4 or H.264/AVC formats. The filter significantly reduces the noisy motion vectors that do not represent a real object movement . The filter analyses the temporal coherence of block motion vectors to determine if they are likely to represent true motion in the recorded scene.

Experiments are performed using the CLEAR metrics for object detection and public available video datasets from CAVIAR, PETS and CLEAR. These experiments demonstrate that the TF outperforms the Vector Median Filter, by providing better object detection accuracy with reduced computational complexity.

The good results obtained by the TF make it suitable as a first step towards implementing systems that aim to detect and track objects from compressed video by using motion vectors. The TF could also be used to improve other techniques based on motion vectors such as Global Motion Estimation (GME) and Motion-Compensated Frame Interpolation (MCFI).

Author Biographies

Ronaldo Carvalho Moura, Monity
Ronaldo Carvalho Moura Junior received the degree of Computer Engineer, in 2004, and the degree of Master of Science in Electronic Engineering and Computer Science, in 2010, both from the Instituto Tecnológico de Aeronáutica, Brazil.

Founder of startup company Monity, since 2005 he works as Researcher and Team Leader for the development of new technologies and products in video processing, embedded systems, and information technology.
Elder Moreira Hemerly, Instituto Tecnológico de Aeronáutica
Elder Moreira Hemerly (S’86–M’89) received the Ph.D. degree in Electrical Engineering from Imperial College, London, U.K., in 1989.

Currently, he is Professor of Control Systems in the Electronics Division, Technological Institute of Aeronautics, Brazil. His current research interests include system identification, adaptive control and signal processing
Adilson Marques da Cunha, Instituto Tecnológico de Aeronáutica
Adilson Marques da Cunha received the Bachelor degree in Air Force Course for Pilots from Brazilian Air Force Academy, in 1970, the Bachelor in Business Administration from Centro de Ensino Unificado de Brasília, in 1979, the degree of Master of Science in Information Systems from Air Force Institute Of Technology- USA, in 1984, and Ph. D. degree in Information Systems from George Washington University-USA, in 1987.

He is Professor in Computer Science Division, Technological Institute of Aeronautics, Brazil. He current research and teaching interests include software engineering; real-time embedded systems; software quality, dependability and safety; CASE tools; CNS-ATM; database system; and artificial intelligence.
Published
09-06-2014
How to Cite
Moura, R., Hemerly, E., & da Cunha, A. (2014). Temporal Motion Vector Filter for Fast Object Detection on Compressed Video. Journal of Communication and Information Systems, 29(1). https://doi.org/10.14209/jcis.2014.1
Section
Regular Papers