Local ternary patterns: Difference between revisions

Revision as of 18:04, 22 October 2012

Latest revision as of 20:12, 26 August 2025

(17 intermediate revisions by 10 users not shown)

Line 1:

{{Short description|Image processing process}}

'''Local ternary patterns''' (LTP) is an extension of the [[Local binary patterns]] introduced by <ref>Xiaoyang Tan and Bill Triggs Enhanced Local Texture Feature Sets for Face Recognition Under Difficult Lighting Conditions, IEEE Transactions on Image Processing, 19(6), pp. 1635-1650, 2010</ref>. Unlike LBP, it does not threshold the pixels into 0 and 1, rather it uses a threshold constant to threshold pixels into three values. Considering k as the threshold constant, c as the value of the center pixel, a neighboring pixel p, the result of threshold is:

{{more citations needed|date=November 2012}}

1 if c>p+k

0 if c>p-k and c<p+k

-1 if c<p-k.

'''Local ternary patterns''' (LTP) isare an extension of the [[Locallocal binary patterns]] introduced by (LBP).<ref>{{Cite journal |last1=Xiaoyang Tan and|last2=Triggs |first2=Bill Triggs|date=June 2010 |title=Enhanced Local Texture Feature Sets for Face Recognition Under Difficult Lighting Conditions, |journal=IEEE Transactions on Image Processing, |volume=19( |issue=6), pp|pages=1635–1650 |doi=10.1109/TIP.2010.2042645 1635|pmid=20172829 |bibcode=2010ITIP...19.1635T |issn=1057-1650,7149|citeseerx=10.1.1.105.3355 2010}}</ref><ref>{{Cite journal |last1=Ji |first1=Luping |last2=Ren |first2=Yan |last3=Pu |first3=Xiaorong |last4=Liu |first4=Guisong |date=2018-07-01 |title=Median local ternary patterns optimized with rotation-invariant uniform-three mapping for noisy texture classification |journal=Pattern Recognition |volume=79 |pages=387–401 |doi=10.1016/j.patcog.2018.02.009|bibcode=2018PatRe..79..387J |s2cid=13691471 |doi-access=free }}</ref> Unlike LBP, it does not threshold the pixels into 0 and 1, rather it uses a threshold constant to threshold pixels[[pixel]]s into three values. Considering k as the threshold constant, c as the value of the center pixel, a neighboring pixel p, the result of threshold is:

In this way, each thresholded pixel has one of the three values. Neighboring pixels are combined after thresholding into a ternary pattern. Computing a histogram of these ternary values will result in a large range, so the ternary pattern is split into two binary patterns. Histograms are concatenated to generate a descriptor double the size of LBP.

:<math>

== Related ==

\begin{cases}

[[Local binary patterns]]

1, & \text{if } p>c+k \\

0, & \text{if } p>c-k \text{ and } p<c+k \\

-1 & \text{if } p<c-k \\

\end{cases}

</math>

In this way, each thresholded pixel has one of the three values. Neighboring pixels are combined after thresholding into a ternary pattern. Computing a [[histogram]] of these [[Ternary numeral system|ternary values]] will result in a large range, so the ternary pattern is split into two binary patterns. Histograms are concatenated to generate a descriptor double the size of LBP.

== RelatedSee also ==

* [[Local binary patterns]]

==References==

{{Reflist}}

[[Category:Computer vision]]

{{comp-sci-stub}}