Pattern Descriptor Research Articles

Two-Dimensional Canonical Correlation Analysis of the Logically Concatenated Cross Binary Pattern for Cross Pose Face Recognition

Face pose recognition is one of the challenging areas in computer vision. Cross-pose change causes the change in the information of face appearance. The maximization of intrasubject correlation helps to widen the intersubject differences which helps further in achieving pose invariance. In this paper, for cross pose recognition, the authors propose to maximize the cross pose correlation by using the logically concatenated cross binary pattern (LC-CBP) descriptor and two dimensional canonical correlation analysis (2DCCA). The LC-CBP descriptor extracts the local texture details of face images with low computation complexity and the 2DCCA explicitly searches for the maximization of the correlated features to retain most informative content. Joint feature consideration via 2DCCA helps in setting up a better correspondence between a discrete set of nonfrontal pose and the frontal pose of the same subject. Experimental results demonstrate the two dimensional canonical correlation LC-CBP descriptor along with intensity values improve the correlation.

Human Epithelial Type-2 Cell Image Classification Using an Artificial Neural Network with Hybrid Descriptors

ABSTRACTAntinuclear antibody (ANA) testing is best performed using the indirect immunofluorescence (IIF) method with human epithelial type-2 (HEp-2) cells as the substrate. IIF is a subjective procedure in which HEp-2 patterns are analyzed manually from the microscope. Therefore, ANA test results greatly rely on the experience and expertise of pathologists. Hence, complete automation of the ANA test is required to avoid incorrect diagnoses. This paper represents an algorithm for the complex HEp-2 cell classification problem. The proposed algorithm used a small hybrid feature set that characterizes the texture and morphology of the HEp-2 cells along with artificial neural network (ANN). The hybrid features were extracted by breaking up the image into eight binary images. The proposed hybrid descriptors were more efficient than the popular co-occurrence matrix descriptor and local binary pattern descriptors for texture analysis. The proposed algorithm was evaluated on the ICPR 2016 IIF HEp-2 cell image dataset. The results indicated that the hybrid descriptor with an ANN approach achieved improved performance, with “96.8%” mean class accuracy.

Improving the classification of rotated images by adding the signal and magnitude information to a local texture descriptor

Texture image classification, especially for images with substantial changes in rotation, illumination, scale and point of view, is a fundamental and challenging problem in the field of computer vision. Natural images acquired under uncontrolled environments have textures with unknown orientation angles. Therefore, it is difficult to identify the same known texture at different acquisition angles. A common solution is image rotation by means of an interpolation technique. However, texture descriptors are not effective enough when two similar textures acquired at different angles are compared. In this work, we propose a simple and efficient image descriptor, called Completed Local Mapped Pattern (CLMP), and apply it to the texture classification of rotated images. This new approach is an improvement over the previously published Local Mapped Pattern (LMP) descriptor because the new approach includes the signal and the magnitude information. This innovation is more discriminating and robust for the description of rotated textures at arbitrary angles. We used two image datasets to validate the proposed descriptor: the Kylberg Sintorn Rotation Dataset and the Brodatz Texture Rotation Dataset. We also introduced a new texture dataset, which contains rotated texture images from Brodatz’s Album. The database contains images of natural textures that have been rotated by both hardware and interpolation methods. We presented an evaluation of the influence of the interpolation method on image rotation, compared with different descriptors in the literature. The experimental results show that our proposed CLMP descriptor outperforms the widely used Completed Local Binary Pattern (CLBP) descriptor and the recently published Sampled Local Mapped Pattern Magnitude (SLMP_M) descriptor. Our results also demonstrate that the choice of interpolation method influences the descriptive capability of each descriptor.

Interactive Visual Exploration of Local Patterns in Large Scatterplot Spaces

AbstractAnalysts often use visualisation techniques like a scatterplot matrix (SPLOM) to explore multivariate datasets. The scatterplots of a SPLOM can help to identify and compare two‐dimensional global patterns. However, local patterns which might only exist within subsets of records are typically much harder to identify and may go unnoticed among larger sets of plots in a SPLOM. This paper explores the notion of local patterns and presents a novel approach to visually select, search for, and compare local patterns in a multivariate dataset. Model‐based and shape‐based pattern descriptors are used to automatically compare local regions in scatterplots to assist in the discovery of similar local patterns. Mechanisms are provided to assess the level of similarity between local patterns and to rank similar patterns effectively. Moreover, a relevance feedback module is used to suggest potentially relevant local patterns to the user. The approach has been implemented in an interactive tool and demonstrated with two real‐world datasets and use cases. It supports the discovery of potentially useful information such as clusters, functional dependencies between variables, and statistical relationships in subsets of data records and dimensions.

Similarity Estimation for Large-Scale Human Action Video Data on Spark

The amount of human action video data is increasing rapidly due to the growth of multimedia data, which increases the problem of how to process the large number of human action videos efficiently. Therefore, we devise a novel approach for human action similarity estimation in the distributed environment. The efficiency of human action similarity estimation depends on feature descriptors. Existing feature descriptors such as Local Binary Pattern and Local Ternary Pattern can only extract texture information but cannot obtain the object shape information. To resolve this, we introduce a new feature descriptor, namely Edge based Local Pattern descriptor (ELP). ELP can extract object shape information besides texture information and ELP can also deal with intensity fluctuations. Moreover, we explore Apache Spark to perform feature extraction in the distributed environment. Finally, we present an empirical scalability evaluation of the task of extracting features from video datasets.

Action recognition based on motion of oriented magnitude patterns and feature selection

Here, the authors introduce a novel system which incorporates the discriminative motion of oriented magnitude patterns (MOMP) descriptor into simple yet efficient techniques. The authors’ descriptor both investigates the relations of the local gradient distributions in neighbours among consecutive image sequences and characterises information changing across different orientations. The proposed system has two main contributions: (i) the authors adopt feature post‐processing principal component analysis followed by vector of locally aggregated descriptors encoding to de‐correlate MOMP descriptor and reduce the dimension in order to speed up the algorithm; (ii) then the authors include the feature selection (i.e. statistical dependency, mutual information, and minimal redundancy maximal relevance) to find out the best feature subset to improve the performance and decrease the computational expense in classification through support vector machine techniques. Experiment results on four data sets, Weizmann (98.4%), KTH (96.3%), UCF Sport (82.0%), and HMDB51 (31.5%), prove the efficiency of the authors’ algorithm.

Face expression recognition using LDN and Dominant Gradient Local Ternary Pattern descriptors

Facial Expression Recognition (FER) is an attractive and demanding problem in human computer interaction. Facial expressions are one of the most dominant and usual means for a human to communicate their feelings and intentions. This paper proposes the Enhanced Modified Decision Based Unsymmetric Trimmed Median Filter (EMDBUTMF) method which removes the noisy pixels from the face image. Also, this paper proposes the Local Directional Number (LDN) pattern, Dominant Gradient Local Ternary Pattern (DGLTP) descriptor for feature extraction and Support Vector Machine (SVM) classifier for classification. The histogram features are extracted from face images using the LDN and DGLTP descriptors. The experimental result shows that the proposed method gives an improved performance of 88% accuracy on two publicly available facial expression datasets JAFFE and CK.

Content-based Image Retrieval for Medicinal Plants Identification using Ensemble of SVM

In order to protect human health, it is very importance to know medicinal plant accurately. The plant identification is not easy and time-consuming task because of diversity. This paper proposes content-based image retrieval (CBIR) for medicinal plant identification using ensemble of Support Vector Machine (SVM). This research developed computerized geometric morphometric method to extract global shape descriptor. We investigated the combination of leaf texture and leaf geometry features based on Fuzzy Local Binary Pattern (FLBP) and global shape descriptor. Weighted Similarity-based Combination method was used to combine the features. We evaluated 660 medicinal leaves images of 66 species. The experiment results show that SVM ensemble of multiple features has good performance with the accuracy of classification is 0.93. CBIR system is developed for medicinal plant identification. The average precision of the CBIR system is 0.913. The proposed method is promising and potential to identify medicinal plant automatically.

Blind Forensics of Images using Higher Order Local Binary Pattern

ABSTRACTExcessive dependencies on digital photographs raise the need of their forensic analysis. When some operation is performed to create fake/forged images then it disturbs the image underlying statistics. To reveal these statistics, image features need to be extracted. The Local Binary Pattern (LBP) descriptor provide good results in texture classification, face recognition, image retrieval, and facial expression recognition, and so forth. It is computationally simple and provides crucial information of image internal statistics. In this article we propose a method for image forgery detection based on higher order LBP texture descriptor. In available literature, higher order pixel analysis gives promising results in various applications that motivate us for proposing a method based on higher order analysis. Higher order analysis provides better correlation information between image pixels that is necessary in image forgery detection. This analysis provide correlation information of near and far pixels in balanced manner, that is, near pixels get more weightage in comparison to far pixels. We assess the efficacy of our proposed work on three publicly available databases and it provides better results in comparison to some of the existing techniques.

Change Detection in Optical Satellite Images Based on Local Binary Similarity Pattern Technique

In this letter, a binary descriptor-based approach is proposed to provide change detection in bitemporal optical satellite images. Multitemporal satellite images suffer from illumination variation. In order to tackle this problem, a variant of a local binary similarity pattern (LBSP) descriptor, which has a good resistance to illumination variation, is proposed. The proposed technique consists of two steps: binary feature vector creation and generation of binary change map using Hamming distance as a similarity metric. To get the binary feature vectors, inter-LBSP, which utilizes the region information across images, is applied on both images. Before applying LBSP, both images are partitioned into overlapping patches. This letter proposes two methods taking different combinations of patches at each pixel position. First one uses the center of one image and neighborhood information of two images, whereas the second one uses the center of two images and neighborhood information of one image. Unlike the conventional LBSP technique, a new threshold calculation technique is used for the generation of binary feature vectors. Experiments are conducted on three data sets acquired by the Landsat satellite, and results show that the proposed technique performs better compared with the earlier reported techniques.

Biologically Inspired Blind Quality Assessment of Tone-Mapped Images

Currently, many tone mapping operators (TMOs) have been provided to compress high dynamic range images to low dynamic range (LDR) images for visualizing them on the common displays. Since quality degradation is inevitably induced by compression, how to evaluate the obtained LDR images is indeed a headache problem. Until now, only a few full reference (FR) image quality assessment metrics have been proposed. However, they highly depend on reference image and neglect human visual system characteristics, hindering the practical applications. In this paper, we propose an effective blind quality assessment method of tone-mapped image without access to reference image. Inspired by that the performance of existing TMOs largely depend on the brightness and chromatic and structural properties of a scene, we evaluate the perceptual quality from the perspective of color information processing in the brain. Specifically, motivated by the physiological and psychological evidence, we simulate the responses of single-opponent (SO) and double-opponent (DO) cells, which play an important role in the processing of the color information. To represent the textural information, we extract three features from gray-level co-occurrence matrix (GLCM) calculated from SO responses. Meanwhile, both GLCM and local binary pattern descriptor are employed to extract texture and structure in the responses of DO cells. All these extracted features and associated subjective ratings are learned to reveal the connection between feature space and human opinion score. Extensive experiments demonstrate that the proposed method outperforms the state-of-the-art blind quality assessment methods and is comparable with the popular FR methods on two recently published tone-mapped image databases.

Support Vector Machine Method to Reduce the Execution Time of Vehicle Plate Recognition System

This research aims to create a vehicle plate detection and recognition system with Cascade Classifier, Support Vector Machine (SVM) and Optical Character Recognition (OCR). Cascade Classifier with Local Binary Patterns (LBP)descriptor is used todetectthe carlicence plate (Coarse Location). SVM is used to reduce plate candidate detection error andthe execution time. Optical Character Recognition (OCR) is used to recognize characters in plates. The system test is performed using 19 video data of moving vehicles at night and rain conditions. Each video has a duration of 30 seconds and contains 4-10 cars per video. The testing results reduce the execution time of vehicle plate recognition systemreached60% with the average accuracy of plate recognition is 61.94%.

Finger Vein Presentation Attack Detection Using Total Variation Decomposition

Finger vein recognition is an emerging biometric technique for personal authentication that has garnered considerable attention in the past decade. Although shown to be effective, recent studies have revealed that finger vein biometrics is also vulnerable to presentation attacks, i.e., printed versions of authorized individual finger vein images can be used to gain access to facilities or services. In this paper, given that both blurriness and the noise distribution are slightly different between real and forged finger vein images, we propose an efficient and robust method for detecting presentation attacks that use forged finger vein images (print artifacts). First, we use total variation regularization to decompose original finger vein images into structure and noise components, which represent the degrees of blurriness and the noise distribution. Second, a block local binary pattern descriptor is used to encode both structure and noise information in the decomposed components. Finally, we use a cascaded support vector machine model for classification, by which finger vein presentation attacks can be effectively detected. To evaluate the performance of our approach, we constructed a new finger vein presentation attack database. Extensive experimental results gleaned from the two finger vein presentation attack databases and a palm vein presentation attack database show that our method clearly outperforms state-of-the-art methods.

Stream data prefetcher for the GPU memory interface

Data caches are often unable to efficiently cope with the massive and simultaneous requests imposed by the SIMT execution model of modern GPUs. While software-aided cache management techniques and scheduling approaches were early considered, efficient prefetching schemes are regarded as the most viable solution to improve the efficiency of the GPU memory subsystem. Accordingly, a new GPU prefetching mechanism is proposed, by extending the stream computing model beyond the actual GPU processing core, thus broadening it toward the memory interface. The proposed prefetcher takes advantage of the available cache management resources and combines a low-profile architecture with a dedicated pattern descriptor specification, which is used to explicitly encode each kernel memory access pattern. The obtained results show that the proposed mechanism increases the L1 data cache hit rate by an average of 61%, resulting in performance speedups as high as 9.2\(\times \) and consequent energy efficiency improvements as high as 11\(\times \).

Logarithmic distance measure with improved local vector pattern for content-based image retrieval

ABSTRACTThe development of image acquisition and data storage mechanisms have empowered the generation of broad image datasets. Under such circumstances, it is essential to promote the relevant information system to handle these collections properly. For this case, one of the most prominent methods is content-based image retrieval (CBIR) system, which can retrieve the required images from the large dataset. This paper proposes the CBIR system based on an informative pattern descriptor and a nonlinear similarity matching measure. The conventional retrieval process determines the distance between the feature library and the query features at the lower order and, subsequently, identifies the similar images with minimum distance. The proposed similarity measure will introduce the degree of similarity in logarithmic scale, and so the range of similar contents can be quantified. The experimental results affirm the beneficial part of PLVP method in providing a more informative and computationally efficient image-retrieving system.

A Novel Technique Based on Visual Words Fusion Analysis of Sparse Features for Effective Content-Based Image Retrieval

Content-based image retrieval (CBIR) is a mechanism that is used to retrieve similar images from an image collection. In this paper, an effective novel technique is introduced to improve the performance of CBIR on the basis of visual words fusion of scale-invariant feature transform (SIFT) and local intensity order pattern (LIOP) descriptors. SIFT performs better on scale changes and on invariant rotations. However, SIFT does not perform better in the case of low contrast and illumination changes within an image, while LIOP performs better in such circumstances. SIFT performs better even at large rotation and scale changes, while LIOP does not perform well in such circumstances. Moreover, SIFT features are invariant to slight distortion as compared to LIOP. The proposed technique is based on the visual words fusion of SIFT and LIOP descriptors which overcomes the aforementioned issues and significantly improves the performance of CBIR. The experimental results of the proposed technique are compared with another proposed novel features fusion technique based on SIFT-LIOP descriptors as well as with the state-of-the-art CBIR techniques. The qualitative and quantitative analysis carried out on three image collections, namely, Corel-A, Corel-B, and Caltech-256, demonstrate the robustness of the proposed technique based on visual words fusion as compared to features fusion and the state-of-the-art CBIR techniques.

Clustering and Jarque-Bera Normality Test to Face Recognition

The normality tests are widely used to estimate the difference between an empirical distribution and the normal distribution. In this paper, we propose to apply the Jarque–Bera normality test (JB-test) in the face recognition field. In order to find out the face areas, which can be approximately represented using a normal distribution, we propose to build a new descriptor based on the JB- test and Local Binary Pattern (LBP) descriptor. To predict whether a pixel belongs to the homogeneous area, Jarque–Bera Local Binary Pattern (JB-LBP) computes the confidence interval basing on the JB- test and uses it later to generate a new pixel grayscale value. The obtained results show that the homogeneous areas pixels are in same way encoded. Therefore, we can easily distinguish them from the peaks areas. Moreover, to reveal that the proposed face representation gives better performances, we combine it with the K-means clustering, the K-Nearest Neighbors (KNN) and the Multi-layers Perceptron (MLP).

Cross-Sensor Fingerprint Matching Method Based on Orientation, Gradient, and Gabor-HoG Descriptors With Score Level Fusion

The fingerprint is one of the oldest and most widely used biometric modality for person identification. Existing automatic fingerprint matching systems perform well when the same sensor is used for both enrollment and verification (regular matching). However, their performance significantly deteriorates when different sensors are used (cross-matching, fingerprint sensor interoperability problem). We propose an automatic fingerprint verification method to solve this problem. It was observed that the discriminative characteristics among fingerprints captured with sensors of different technology and interaction types are ridge orientations, minutiae, and local multi-scale ridge structures around minutiae. To encode this information, we propose two minutiae-based descriptors: histograms of gradients obtained using a bank of Gabor filters and binary gradient pattern descriptors, which encode multi-scale local ridge patterns around minutiae. In addition, an orientation descriptor is proposed, which compensates for the spurious and missing minutiae problem. The scores from the three descriptors are fused using a weighted sum rule, which scales each score according to its verification performance. Extensive experiments were conducted using two public domain benchmark databases (FingerPass and Multi-Sensor Optical and Latent Fingerprint) to show the effectiveness of the proposed system. The results showed that the proposed system significantly outperforms the state-of-the-art methods based on minutia cylinder-code (MCC), MCC with scale, VeriFinger—a commercial SDK, and a thin-plate spline model.

Pollen Image Recognition Based on DGDB-LBP Descriptor

In this paper, we propose DGDB-LBP, a local binary pattern descriptor based on the pixel blocks in the dominant gradient direction. Differing from traditional LBP and its variants, DGDB-LBP encodes by comparing the main gradient magnitude of each block rather than the single pixel value or the average of pixel blocks, in doing so, it reduces the influence of noise on pollen images and eliminates redundant and non-informative features. In order to fully describe the texture features of pollen images and analyze it under multi-scales, we propose a new sampling strategy, which uses three types of operators to extract the radial, angular and multiple texture features under different scales. Considering that the pollen images have some degree of rotation under the microscope, we propose the adaptive encoding direction, which is determined by the texture distribution of local region. Experimental results on the Pollenmonitor dataset show that the average correct recognition rate of our method is superior to other pollen recognition methods in recent years.

High Dimensional Feature for Hyperspectral Image Classification

Making a high dimensional (e.g., 100k-dim) feature for hyperspectral image classification seems not a good idea because it will bring difficulties on consequent training, computation, and storage. In this paper, we study the performance of a high-dimensional feature by texture feature. The texture feature based on multi-local binary pattern descriptor, can achieve significant improvements over both its tradition version and the one we proposed in our previous work. We also make the high-dimensional feature practical, we employ the PCA method for dimension reduction and support vector machine for hyperspectral image classification. The two real hyperspectral image datasets are employed. Our experimental results with real hyperspectral images indicate that the high dimensional feature can enhance the classification accuracy than some low dimensional.