Efficient disparity estimation from stereo images using hybrid-guided image filter

ABSTRACT

Stereo vision process involves capturing the pictures from a camera of the same scene from at least two different locations and calculating the three-dimensional information. Conventionally, these two versions of snapshots are called left and right views which yield the depth information of an object upon relative comparison of its location in two views. Although the stereo image and its applications are becoming increasingly prevalent, there has been very limited research on disparity estimation from stereo images. Most of the existing techniques suffer from the gradient reversal artefacts issue. Therefore, to handle this issue, we have proposed a hybrid-guided image filter for improving the disparity estimation from stereo images. The hybrid filter utilizes the features of guided image filter and Bayesian non-local means with edge aware constraint. Maximum likelihood and local area homogeneity analysis are used to generate the guidance image for the proposed filter. To enhance the quality of disparity estimation from stereo images, segmentation is also done using the modified mean shift technique. Experimental results show that the proposed technique can efficiently estimate the depth maps over the available techniques. One-way ANOVA analysis on experimental results validates that the hybrid filter-based stereo matching outperforms consistently over the state-of-art approaches.

KEYWORDS:

• Stereo matching
• guided image filter
• Bayesian non-local means
• winner-take-all
• modified mean shift

Disclosure statement

No potential conflict of interest was reported by the authors.

Notes on contributors

Khushboo Jain received her B.Tech degree in CSE from Rajasthan Technical University, India, in 2013 and M. Tech. Degree in Computer Science and Applications from Thapar University, Patiala, in 2016. She currently works as a scientist at Bharat Electronics Ltd. India. Her research interests include 3-D modelling and computer vision.

Husanbir Singh Pannu is working as a lecturer in Thapar University, Patiala, India. He completed his Ph.D. from University of North Texas, USA, and MS from California State University, East Bay, USA. His research interests include machine learning, optimization, and image processing.

Kuldeep Singh received his B.E. degree in Electronics & Communication from Government Engineering College, Ajmer, India, in 2004 and M. Tech. degree in Signal Processing from Delhi University in 2006 and Ph.D. degree in Computer Vision from Delhi Technological University, New Delhi, India, in 2016. He currently works as a scientist at Bharat Electronics Ltd. India. His research interest includes deep learning-based computer vision, medical imaging, crowd behavior analysis, human action/activity recognition, image quality improvement and sparse representation. He has more than 10 publications in renowned journals of computer vision area. He is also a reviewer of various Elsevier, IEEE and IET journals.

Avleen Kaur Malhi is currently working as Assistant Professor in Thapar University, Patiala. She completed her Ph.D. in the area of security of VANETs in the Department of Computer Science and Engineering at Thapar University, Patiala, Punjab, India, in 2016 and received M.E. (CSE) from Thapar University, Patiala, Punjab, India, in 2012. Her research interests include vehicular networks, information security, and machine learning.

ORCID

Kuldeep Singh http://orcid.org/0000-0002-2350-1700

Note on Images

All Middlebury Stereo Datasets can be obtained from the following link: http://vision.middlebury.edu/stereo/data/