Periodicity and pattern size detection for quality improvement of the patch-based structural texture synthesis
by Charushila Vijay Rane; Sandip Raosaheb Patil
International Journal of Applied Pattern Recognition (IJAPR), Vol. 6, No. 4, 2021

Abstract: The texture is the surface property that contains the repetition of patterns. One iteration of a pattern is called the texel. The regularity in the texels and texel's placement defines the texture classification as regular/structural or non-regular/random texture. The pattern size as well as the periodicity of textures is derived from the Brodatz album. The correlation between two adjacent patches is calculated instead of the autocorrelation of the whole texture image. The patch size is varied each time. If patch size matches with pattern size or its multiples, the correlation value is at peak. Repeated occurrences of correlation peaks, variation in peak correlation values, and strength of peak correlation values decide the periodicity of texture. Out of 112 textures of Brodatz album, 60 textures have a periodicity of more than 60%. The synthesised structural texture's quality is much improved when patches of pattern size are used instead of using patches of random size. The structural similarity of synthesised texture is increased from 0.0947 to 0.4142.

Online publication date: Thu, 11-Nov-2021

The full text of this article is only available to individual subscribers or to users at subscribing institutions.

 
Existing subscribers:
Go to Inderscience Online Journals to access the Full Text of this article.

Pay per view:
If you are not a subscriber and you just want to read the full contents of this article, buy online access here.

Complimentary Subscribers, Editors or Members of the Editorial Board of the International Journal of Applied Pattern Recognition (IJAPR):
Login with your Inderscience username and password:

Username:        Password:          Forgotten your password?

Want to subscribe?
A subscription gives you complete access to all articles in the current issue, as well as to all articles in the previous three years (where applicable). See our Orders page to subscribe.

If you still need assistance, please email subs@inderscience.com