Texture discrimination with multidimensional distributions of signed gray-level differences

Ojala, Timo; Valkealahti, Kimmo; Oja, Erkki; Pietikäinen, Matti

doi:10.1016/s0031-3203(00)00010-8

Cited by 221 publications

(131 citation statements)

References 15 publications

Supporting

Mentioning

126

Contrasting

Unclassified

Order By: Relevance

“…5. For Brodatz32 [16], it comprises 2,048 samples, with 64 samples in each of the 32 texture categories [6,16,22]. UIUC has 25 classes with images under uncontrolled illumination, albedo variations, 3D shape, as well as a mixture of both.…”

Section: Dataset and Set-upsmentioning

confidence: 99%

RLBP: Robust Local Binary Pattern

Chen

Kellokumpu

Zhao

et al. 2013

Procedings of the British Machine Vision Conference 2013

View full text Add to dashboard Cite

In this paper, we propose a simple and robust local descriptor, called the robust local binary pattern (RLBP). The local binary pattern (LBP) works very successfully in many domains, such as texture classification, human detection and face recognition. However, an issue of LBP is that it is not so robust to the noise present in the image. We improve the robustness of LBP by changing the coding bit of LBP. Experimental results on the Brodatz and UIUC texture databases show that RLBP impressively outperforms the other widely used descriptors (e.g., SIFT, Gabor, MR8 and LBP) and other variants of LBP (e.g., completed LBP), especially when we add noise in the images. In addition, experimental results on human face recognition also show a promising performance comparable to the best known results on the Face Recognition Grand Challenge (FRGC) face dataset. IntroductionRecently, many sparse and dense descriptors (e.g., SIFT, Gabor, MR8 and LBP) have been proposed for different kinds of applications. There are several studies to evaluate their performance, e.g., [13,14]. LBP [15] is perhaps the best performing dense descriptor and it has been widely used in various applications, such as texture classification, human detection and face recognition [18]. It has been proven to be highly discriminative and its key advantages, namely its invariance to monotonic gray level changes and computational efficiency, make it suitable for demanding image analysis tasks.However, one issue of LBP is that it is not so robust to the noise present in images when the gray-level changes resulting from the noise are not monotonic, even if the changes are not significant [2]. To this end, we propose a new descriptor based on LBP, i.e., robust local binary pattern (RLBP). The idea is to locate the possible bit in LBP pattern changed by the noise and then revise the changed bit of the LBP pattern. The idea is very simple, but it works very well. For example, the performance of LBP decreases significantly when we add white Gauss noise in the Brodatz texture dataset [1]. However, the performance of RLBP almost does not change. We also add noise in UIUC texture [7] and FRGC face datasets [17] to testify the performance of RLBP.

show abstract

Section: Dataset and Set-upsmentioning

confidence: 99%

RLBP: Robust Local Binary Pattern

Chen

Kellokumpu

Zhao

et al. 2013

Procedings of the British Machine Vision Conference 2013

View full text Add to dashboard Cite

show abstract

“…Gray scale invariance is also necessary if the gray scale properties of the training and testing data differ. This was clearly demonstrated in our recent study [9] on supervised texture segmentation with the same the image set that was used by Randen and Husoy in their recent extensive comparative study [12]. However, real world textures with a large tactile dimension can also exhibit non-monotonic intensity changes, e.g.…”

Section: Discussionmentioning

confidence: 62%

“…Since LBP P,R riu2 has a completely defined set of discrete output values (0 -> P+1), no additional binning procedure is required, but the operator outputs are directly accumulated into a histogram of P+2 bins. Variance measure VAR P,R has a continuous-valued output, hence quantization of its feature space is needed, together with the selection of an appropriate value for B [9].…”

Section: Nonparametric Classification Principlementioning

confidence: 99%

A Generalized Local Binary Pattern Operator for Multiresolution Gray Scale and Rotation Invariant Texture Classification

Ojala

Pietikäinen

Mäenpää

2001

Lecture Notes in Computer Science

Self Cite

233

113

View full text Add to dashboard Cite

Abstract. This paper presents generalizations to the gray scale and rotation invariant texture classification method based on local binary patterns that we have recently introduced. We derive a generalized presentation that allows for realizing a gray scale and rotation invariant LBP operator for any quantization of the angular space and for any spatial resolution, and present a method for combining multiple operators for multiresolution analysis. The proposed approach is very robust in terms of gray scale variations, since the operator is by definition invariant against any monotonic transformation of the gray scale. Another advantage is computational simplicity, as the operator can be realized with a few operations in a small neighborhood and a lookup table. Excellent experimental results obtained in a true problem of rotation invariance, where the classifier is trained at one particular rotation angle and tested with samples from other rotation angles, demonstrate that good discrimination can be achieved with the occurrence statistics of simple rotation invariant local binary patterns. These operators characterize the spatial configuration of local image texture and the performance can be further improved by combining them with rotation invariant variance measures that characterize the contrast of local image texture. The joint distributions of these orthogonal measures are shown to be very powerful tools for rotation invariant texture analysis. nonparametric texture analysis distribution histogram Brodatz contrast

show abstract

“…These methods have been extensively analysed and used in many studies and applications in image processing and computer vision (Zhang et al, 2002). Some studies have also pointed out that these texture features can perform better than the TGMRF features (Ojala et al, 2001;Hadjidemetriou et al, 2003;Pietikäinen et al, 2000;Liu and Wang, 2003). Therefore, rotational invariant uniform local binary patterns (LBP) (Ojala et al, 2002) and spectral histograms (SH) (Liu and Wang, 2003) are employed in our study for the comparison.…”

Section: Comparison To Other Texture Descriptorsmentioning

confidence: 99%

“…TGMRF features describe spatial pixel dependencies which is a primary characteristic associated with texture. However, these features ignore some important structural and statistical information about the texture and have performed poorly (Ojala et al, 2001;Hadjidemetriou et al, 2003;Pietikäinen et al, 2000;Petrou and Sevilla, 2006;Liu and Wang, 2003). Therefore in recent work, we proposed Local Parameter Histogram (LPH) descriptor which is an improved texture descriptor demonstrating significant improvement in characterizing texture compared to the TGMRF descriptors (Dharmagunawardhana et al, 2014b) .…”

Section: Introductionmentioning

confidence: 99%