Comparison of Local Feature Extraction Paradigms Applied to Visual SLAM

Authors

  • Víctor R. López López Instituto Tecnológico de Tijuana; Departamento de Ingeniería Eléctrica y Electrónica, Tijuana, BC
  • Leonardo Trujillo Instituto Tecnológico de Tijuana; Departamento de Ingeniería Eléctrica y Electrónica, Tijuana
  • Pierrick Legrand Universite´ of Bordeaux, CQFD Team, INRIA Bordeaux, IMB, Talence, France
  • Victor H. Díaz Ramírez Instituto Politécnico Nacional, CITEDI
  • Gustavo Olague CICESE, EvoVisión Group, Applied Physics Division, Ensenada, Baja California

DOI:

https://doi.org/10.13053/cys-20-4-2500

Keywords:

Local features, genetic programming, composite correlation filter, SLAM.

Abstract

The detection and description of locally salient regions is one of the most widely used low-level processes in modern computer vision systems. The general approach relies on the detection of stable and invariant image features that can be uniquely characterized using compact descriptors. Many detection and description algorithms have been proposed, most of them derived using different assumptions or problem models. This work presents a comparison of different approaches towards the feature extraction problem, namely: (1) standard computer vision techniques; (2) automatic synthesis techniques based on genetic programming (GP); and (3) a new local descriptor based on composite correlation filtering, proposed for the first time in this paper. The considered methods are evaluated on a difficult real-world problem, vision-based simultaneous localization and mapping (SLAM). Using three experimental scenarios,   results indicate that the GP-based methods and the correlation filtering techniques outperform widely used computer vision algorithms such as the Harris and Shi-Tomasi detectors and the Speeded Up Robust Features descriptor.

Author Biographies

Víctor R. López López, Instituto Tecnológico de Tijuana; Departamento de Ingeniería Eléctrica y Electrónica, Tijuana, BC

Received a degree in Electronic Engineering (2012) and a Masters in Sciences of Engineering (2015) from the Instituto Tecnológico  de Tijuana, México. He is currently studying a PhD in Sciences of Engineering in the Instituto Tecnológico de Tijuana, México, where his research interests are Genetic Programming (GP), Computer Vision (CV) and Human-Computer interface (HCI). He is also a current member of the TREE-LAB research group.

Leonardo Trujillo, Instituto Tecnológico de Tijuana; Departamento de Ingeniería Eléctrica y Electrónica, Tijuana

Received a degree in Electronic Engineering (2002) and a Masters in Computer Science (2004) from the Instituto Tecnológico de Tijuana, México. He also received a doctorate in Computer Science from the CICESE research center, in Ensenada México (2008), developing Genetic Programming (GP) applications for Computer Vision problems, focusing on feature extraction and image description. He is currently professor at Instituto Tecnológico de Tijuana, México (ITT), where he is currently President of the Master Program, head of the Cybernetics research group and is the head researcher of the TREE-LAB research group.

Pierrick Legrand, Universite´ of Bordeaux, CQFD Team, INRIA Bordeaux, IMB, Talence, France

Received his PhD in applied mathematics (multifractal analysis, signal processing) from Ecole centrale de Nantes and from Nantes university in December 2004. During his thesis, under the supervision of Jacques Levy-Vehel, he developed interpolation and denoising methods based on Hölderian regularity and  wavelets, proving that these methods reach an optimal rate of convergence. Moreover, during his thesis he was in charge of the development of FracLab, a free Matlab toolbox. In 2005 he received a post-doctoral position shared between the Evovision group at CICESE research center (Ensenada, México) and INRIA COMPLEX Team (Rocquencourt, France). In 2006, a second post-doctoral position allowed Pierrick Legrand to develop the first genetic algorithm running on a pocket PC. On September 2006, he became associate professor at the university of Bordeaux (UFR Sciences and Modelisation) and researcher at the IMB (Institut de Mathématiques de Bordeaux, UMR CNRS 5251), both of which are positions he currently holds. In 2008, Pierrick Legrand, started to develop and apply signal processing technics and evolutionary computation methods to the analysis of EEG signals. Since 2010 he is also a researcher at INRIA, the same year he was awarded two best-paper awards from the leading conference in evolutionary computation, GECCO 2010.

Victor H. Díaz Ramírez, Instituto Politécnico Nacional, CITEDI

Obtained his MS degree in electronics engineering from Instituto Tecnológico de Mexicali in 2003 and his PhD in computer science from Centro de Investigación  Científica y de Educación Superior de Ensenada (CICESE), México, in 2007. He is now a professor at Instituto Politécnico Nacional, México. His research interests include signal and image processing, pattern recognition, and opto-digital correlators.

Gustavo Olague, CICESE, EvoVisión Group, Applied Physics Division, Ensenada, Baja California

Received his Ph.D. in Computer Vision, Graphics and Robotics from INPG (Institut Polytechnique de Grenoble) and INRIA (Institut National de Recherche en Informatique et Automatique) in France. He is a Professor in the Dept. of Computer Science at CICESE (Centro de Investigación on Científica y de Educación Superior de Ensenada) in México, and the Director of its EvoVisión Research Team. He is also an Adjoint Professor of Engineering at UACH (Universidad Autónoma de Chihuahua). He has authored over 100 conference proceedings papers and journal articles, he coedited two special issues in Pattern Recognition Letters and Evolutionary Computation, and he served as cochair of the Real-World Applications track at the main international evolutionary computing conference, GECCO (ACM SIGEVO Genetic and Evolutionary Computation Conference). Prof. Olague has received numerous distinctions, among them the Talbert Abrams Award presented by the American Society for Photogrammetry and Remote Sensing (ASPRS) for authorship and recording of current and historical engineering and scientific developments in photogrammetry; Best Paper awards at major conferences such as GECCO, EvoIASP (European Workshop on Evolutionary Computation in Image Analysis, Signal Processing and attern Recognition) and EvoHOT (European Workshop on Evolutionary Hardware Optimization); and twice the Bronze Medal at the Humies (GECCO award for Human-Competitive results produced by genetic and evolutionary computation). His main research interests are evolutionary computing computer vision. He is author of the book Evolutionary Computer Vision published by Springer

Downloads

Published

2016-12-18

Issue

Vol. 20 No. 4 (2016): Thematic Issue: Research Advances and Applications of Evolutionary Computation (Guest Editors: H.J. Escalante, M. Graff)

Section

Articles of the Thematic Issue

License 

Hereby I transfer exclusively to the Journal "Computación y 
Sistemas", published by the Computing Research Center (CIC-IPN),
the Copyright of the aforementioned paper. I also accept that these 
rights will not be transferred to any other publication, in any other 
format, language or other existing means of developing.
I certify that the paper has not been previously disclosed or simultaneo
usly submitted to any other publication, and that it does not contain 
material whose publication would violate the Copyright or other 
proprietary rights of any person, company or institution. I certify that 
I have the permission from the institution or company where I work or 
study to publish this work.
The representative author accepts the responsibility for the publication
of this paper on behalf of each and every one of the authors. 
This transfer is subject to the following conditions:
  • The authors retain all ownership rights (such as patent rights) of this work, except for the publishing rights transferred to the CIC, through this document.
  • Authors retain the right to publish the work in whole or in part in any book they are the authors or publishers. They can also make use of this work in conferences, courses, personal web pages, and so on.
  • Authors may include working as part of his thesis, for non-profit distribution only.