Lbp based edge-texture features for object recoginition
-
Upload
igeeks-technologies -
Category
Education
-
view
128 -
download
1
description
Transcript of Lbp based edge-texture features for object recoginition
LBP-Based Edge-Texture Features forObject RecognitionAmit Satpathy, Member, IEEE, Xudong Jiang, Senior Member, IEEE, and How-Lung Eng, Member, IEEE
Abstract
This paper proposes two sets of novel edge-texture features, Discriminative Robust Local Binary Pattern (DRLBP) and Ternary Pattern (DRLTP), for object recognition. By investigating the limitations of Local Binary Pattern (LBP), Local Ternary Pattern (LTP) and Robust LBP (RLBP), DRLBP and DRLTP are proposed as new features. They solve the problem of discrimination between a bright object against a dark background and vice-versa inherent in LBP and LTP. DRLBP also resolves the problem of RLBP whereby LBP codes and their complements in the same block are mapped to the same code. Furthermore, the proposed features retain contrast information necessary for proper representation of object contours that LBP, LTP, and RLBP discard. Our proposed features are tested on seven challenging data sets: INRIA Human, Caltech Pedestrian, UIUC Car, Caltech 101, Caltech 256, Brodatz, and KTH-TIPS2- a. Results demonstrate that the proposed features outperform the compared approaches on most data sets.
Existing System
Object recognition features are categorized into two groups sparse and dense representations. For Sparse feature representations, interest-point detectors are used to identify structures such as corners and blobs on the object. A feature is created for the image patch around each point. Popular feature representations include Scale-Invariant Feature Transform (SIFT) like Speeded Up Robust Feature, Local Steering Kernel, Principal Curvature-Based Regions, Region Self-Similarity features. For Dense feature representations, which are extracted at fixed locations densely in a detection window, are gaining popularity as they describe objects richly compared to sparse feature representations. Various feature representations such as Wavelet, Haar-like features, Histogram of Oriented Gradients (HOG), Extended Histogram of Gradients, feature Context.
Disadvantages of Existing System
The present system differentiates a bright object against a dark background and vice versa. This increases the object intra-class variations which is undesirable for most object recognitions.
Different objects have different shapes and textures. It is therefore desirable to represent objects using both texture and edge information.
The present system only captures texture information. Object contours, which also contain discriminatory information, tend to be situated in strong contrast regions, are discarded. Therefore, by discarding contrast information, contours may not be effectively represented.
Proposed System
In this paper, we propose two sets of novel edge-texture features, Discriminative Robust LBP (DRLBP) and LTP. The proposed features solve the issues of LBP, LTP and RLBP. They alleviate the intensity reversal problem of object and background. Furthermore, DRLBP discriminates local structures that RLBP misrepresent. In addition, the proposed features retain the contrast information of image patterns. They contain both edge and texture information which is desirable for object recognition. Results indicate that the proposed features outperform LBP,LTP and RLBP and perform better than other approaches in comparison on datasets.
Advantages of Proposed System
The proposed features retain the contrast information of image patterns. They contain both edge and texture information which is desirable for object recognition.
The proposed features solve the issues of LBP, LTP and RLBP. They alleviate the intensity reversal problem of object and background. Furthermore, DRLBP discriminates local structures that RLBP misrepresent.
Modules
The proposed system has the following modules:
Local Binary Pattern (LBP)
Local Ternary Pattern (LTP)
Robust Local Binary Pattern (RLBP)
Robust Local Ternary Pattern (RLTP)
Kernel Algorithm
Modules Description
Local Binary Pattern (LBP)LBP is a type of feature used for classification in computer vision. LBP is the particular case of the Texture Spectrum model. It has to be a powerful feature for texture classification; it has further been determined that when LBP is combined with the Histogram of oriented gradients (HOG) descriptor, it improves the detection performance considerably on some datasets.
Local Ternary Pattern (LTP)
(LTP) are an extension of Local binary patterns (LBP). Unlike LBP, it does not threshold the pixels into 0 and 1. 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.
Robust Local Binary Pattern (RLBP)
Robust Local Binary Patterns (RLBP) works very successfully in many domains, such as texture classification, human detection and face recognition. An issue of LBP is that is not so robust to the noise present in the image. We improve the robustness of LBP by changing the coding bit of LBP. RLBP impressively outperforms the other widely used descriptors.
Robust Local Ternary Pattern (RLTP)
Robust Local Ternary Patterns (RLTP) works very successfully in many domains, such as texture classification, human detection and face recognition. An issue of LTP is that is not so robust to the noise present in the image. We improve the robustness of LTP by changing the coding bit of LTP. RLTP impressively outperforms the other widely used descriptors.
Kernel Algorithm
Kernel Algorithms are a class of algorithms for pattern analysis. The general task of pattern analysis is to find and study general types of relations in datasets. For many of these tasks, data have to be represented as feature vectors, but kernel methods replace this representation by similarities to other data points. Kernel methods owe their name to the use of kernel functions, which enable them to operate in a high-dimensional, implicit feature space without ever computing the coordinates of the data in that space, but rather by simply computing the inner products between the images of all pairs of data in the feature space. This operation is often computationally cheaper than the explicit computation of the coordinates. This approach is called the kernel trick. Kernel functions have been introduced for sequence data, graphs, text, images, as well as vectors.
Conclusion
This paper proposes 2 sets of novel edge-texture features, Discriminative Robust Local Binary Pattern (DRLBP) and Ternary Pattern (DRLTP), for object recognition. The limitations of existing texture features, Local Binary Pattern (LBP), Local Ternary Pattern (LTP) and RobustLBP (RLBP), for object recognition are analyzed. LBP and LTP differentiate a bright object against a dark background and vice-versa. This differentiation makes the object intra-class variations larger. RLBP solves the LBP problem by choosing the minimum of a LBP code and its complement. However, RLBP maps LBP codes and their complement in the same block to the same value. This causes some structures to be misrepresented. Furthermore, LBP, LTP and RLBP discard contrast information. This is not desired as object texture and contour both contain discriminative information. By capturing only the texture information, the contour is not effectively represented. The new features, DRLBP and DRLTP, are proposed by analyzing the weaknesses of LBP, LTP and RLBP. They alleviate the problems of LBP, LTP and RLBP by considering both the weighted sum and absolute difference of the bins of the LBP and LTP codes with their respective complement/inverted codes. The new features are robust to image variations caused by the intensity inversion and are discriminative to the image structures within the histogram block. We present results of the proposed features on 7 data sets and compare them with several methods for object recognition. Results demonstrate that the proposed features outperform the compared recognition approaches on most data sets.
Minimum Hardware Configuration of the Proposed System
Processor : Intel/AMD
Hard Disk : 40 GB
Monitor : 14’ Colour Monitor
Keyboard : Standard Keyboard
Mouse : Optical Mouse
RAM : 512 MB
Software Configuration of the Proposed System Operating system : Windows 7 onwards 32-bit
Coding Language : ASP.Net with C#
Database : SQL Server 2008
References
S. Agarwal, A. Awan, and D. Roth, “Learning to detect objects in images via a sparse, part-based representation,” IEEE Trans. Pattern Anal. Mach. Intell., vol. 26, no. 11, pp. 1475–1490, Nov. 2004.
T. Ahonen, A. Hadid, and M. Pietikainen, “Face description with local binary patterns: Application to face recognition,” IEEE Trans. Pattern Anal. Mach. Intell., vol. 28, no. 12, pp. 2037–2041, Dec. 2006.
H. Bay, A. Ess, T. Tuytelaars, and L. J. V. Gool, “Speeded-up robust features (surf),” Comput. Vis. Image Understand., vol. 110, no. 3, pp. 346–359, 2008.
O. Boiman, E. Shechtman, and M. Irani, “In defense of nearest-neighbor based image classification,” in Proc. IEEE Int. Conf. Comput. Vis. Pattern Recognit., Jun. 2008, pp. 1–8.
P. Brodatz, Textures: A Photographic Album for Artists and Designers. New York, NY, USA: Dover Publications, Aug. 1999.