By Topic

Shape from shading using wavelets and weighted smoothness constraints

Sign In

Cookies must be enabled to login.After enabling cookies , please use refresh or reload or ctrl+f5 on the browser for the login options.

Formats Non-Member Member
$31 $31
Learn how you can qualify for the best price for this item!
Become an IEEE Member or Subscribe to
IEEE Xplore for exclusive pricing!
close button

puzzle piece

IEEE membership options for an individual and IEEE Xplore subscriptions for an organization offer the most affordable access to essential journal articles, conference papers, standards, eBooks, and eLearning courses.

Learn more about:

IEEE membership

IEEE Xplore subscriptions

2 Author(s)
Chen, D. ; Sch. of Inf. Syst., Comput. & Math., Brunel Univ., Uxbridge, UK ; Dong, F.

A new method that allows capturing shapes from an input image using an optimisation-based approach is presented. An objective function is designed by introducing two terms: the first term is used to minimise the difference between the shading of the reconstructed shape and the input image, and the second term is to apply smoothness constraints to the reconstructed shape. To achieve shape reconstruction in high quality, the authors propose weighted smoothness constraint, which is designed to be anti-proportional to the intensity gradients in the input image. Under this constraint, flat image areas make more contribution towards the smoothness of the reconstructed shape, while the fine details from the image areas with large intensity gradients are preserved in the reconstructed result. Given the objective function, wavelets are used to obtain the solution effectively. Since wavelets accurately preserve high-frequency data, they can be used to solve the objective function with the advantage of allowing for a good recovery of fine details from the input image. The authors have chosen to use the Daubechies wavelets, which are orthonormal and compactly supported. Here the formulation of the algorithm based on the mathematical details is provided. Finally, the authors present experimental results on a number of different images and compare them against some well-known methods and ground truth (where available). The comparison shows that the method is effective and offers good results.

Published in:

Computer Vision, IET  (Volume:4 ,  Issue: 1 )