Real Image Database Research Articles

Computer Aided Diagnosis System-A Decision Support System for Clinical Diagnosis of Brain Tumours

The iso, hypo or hyper intensity, similarity of shape, size and location complicates the identification of brain tumors. Therefore, an adequate Computer Aided Diagnosis (CAD) system is designed for classification of brain tumor for assisting inexperience radiologists in diagnosis process. A multifarious database of real post contrast T1-weighted MR images from 10 patients has been taken. This database consists of primary brain tumors namely Meningioma (MENI- class 1), Astrocytoma (AST- class 2), and Normal brain regions (NORM- class 3). The region of interest(s) (ROIs) of size 20 x 20 is extracted by the radiologists from each image in the database. A total of 371 texture and intensity features are extracted from these ROI(s). An Artificial Neural Network (ANN) is used to classify these three classes as it shows better classification results on multivariate non-linear, complicated, rule based domains, and decision making domains. It is being observed that ANN provides much accurate results in terms of individual classification accuracy and overall classification accuracy. The four discrete experiments have been performed. Initially, the experiment was performed by extracting 263 features and an overall classification accuracy 78.10% is achieved, however, it was noticed that MENI (class-1) was highly misclassified with AST (class-2). Further, to improve the overall classification accuracy and individual classification accuracy specifically for MENI (class-1), LAWs textural energy measures (LTEM) are added in the feature bank (263+108=371). An individual class accuracy of 91.40% is obtained for MENI (class-1), 91.43% for AST (class-2), 94.29% for NORM (class-3) and an overall classification accuracy of 92.43% is achieved. The results are calculated with and without addition of LTEM feature with Principle component analysis (PCA)-ANN. LTEM-PCA-ANN approach improved results with an overall accuracy of 93.34%. The texture patterns obtained were clear enough to differentiate between MENI (class-1) and AST (class-2) despite of necrotic and cystic component and location and size of tumor. LTEM detected fundamental texture properties such as level, edge, spot, wave and ripple in both horizontal and vertical directions which boosted the texture energy.

Salient object detection of social images based on semantic tag context

Salient object detection is an important process for machines to understand visual contents as humans. Typically, most previous studies on salient object detection infer salient map by only using the visual features. In this paper, we propose a new paradigm on salient object detection, which aims at producing more reliable results by mining the context information from user's annotated tags. To address this problem, we firstly construct a large scale salient object dataset, which includes 5429 images from the NUSWIDE dataset (a real world web image database from National University of Singapore) with tag information and accurate human-labeled masks. Moreover, a specialised conditional random field (CRF) model is also proposed which takes account of both tag contexts and appearance cues. Extensive experiments show the effectiveness and superiority of our model which can accurately locate the salient objects.

Segmentation of carbon nanotube images through an artificial neural network

Segmentation of carbon nanotube images is an important task for nanotechnology. The segmentation stage determines the accuracy of the measurement process of nanotube when assessing the quality of nanomaterials. In this work, we propose two segmentation algorithms for carbon nanotube images. Each algorithm includes three stages: preprocessing, segmentation and postprocessing. The first one is applied on images from scanning electron microscopy and employs a matched filter bank in the preprocessing step followed by a neural network in the segmenting phase. The second algorithm uses the Perona---Malik filter for enhancing the nanotube information. The segmentation phase is composed of the relaxed Otsu's threshold and an artificial neural network. This algorithm is applied on images from transmission electron microscopy. The postprocessing stage, for both algorithms, is based on mathematical morphology. The performance of the proposed algorithms is numerically evaluated by using real image databases, manually segmented by an expert. The algorithm for segmentation of scanning electron microscopy achieved 92.74% of overall accuracy, while the algorithm for segmentation of transmission electron microscopy obtained an accuracy of 73.99% if the whole image is considered. A performance improvement is accomplished if only the region of interest is segmented, arriving to 84.19% of overall accuracy.

SYSTEMATIZATION OF SPACE OF STRUCTURAL FEATURES BASED ON SELF-LEARNING METHODS FOR EFFECTIVE IMAGE RECOGNITION

The work deals with issues of clustering sets of characteristic features of images. For the construction of array of the characteristic features is used method Speeded Up Robust Features. Implemented algorithms for clustering structural descriptions of images on the basis of a self-organizing Kohonen neural network and method of grouping the difference. The object of the research are clustering methods which applied to the set of structural features. The aim is to construct a vector representations of descriptions based on clustering, which increases the speed of recognition. The subject of research is systematization a set of structural features of visual objects. Discussing the results of the application of clustering methods for structural descriptions of images in the form of sets of characteristic features to improve the performance of visual recognition of objects. For systematization and compression the feature space proposed to carry out self-study using the methods of differential grouping and Kohonen networks. The simulation and experimental study of clustering methods on examples of specific sets of characteristic features were done. The research results proves the possibility of effective representation of the descriptions in the form of a vector with integer elements. This approach can be used to solve problems of recognition and retrieval of images. As a result compact vector description of etalon images is built, quantitative estimates of clustering error are estimated, efficiency of proposed method during processing of real image database is confirmed.

Defect detection on patterned fabrics using texture periodicity and the coordinated clusters representation

Patterned fabrics may be regarded as periodic textures, which are defined as the regular tessellation of a primitive unit. A patterned fabric is considered as defective when a primitive unit is different from the others. In this paper, we propose a one-class classifier that uses Reduced Coordinated Cluster Representation (RCCR) as features. In the training step, the size of the primitive unit of defect-free fabrics is automatically estimated using a texture periodicity algorithm. After that, the fabrics are split into samples of one unit and their local structure is learnt with the RCCR features in a one-class classifier. During the test step, defective and non-defective fabrics are also split into samples and are analyzed unit by unit. If the features of a given unit do not satisfy the classification criterion, it is considered to be a defect. Among the advantages of the RCCR is that it represents structural information of textures in a low-dimensional feature space with high discrimination performance. Results from experiments on an extensive database of real fabric images show that our method yields accurate detections, outperforming other state-of-the-art algorithms.

Unsupervised fuzzy-wavelet framework for coastal polynya detection in synthetic aperture radar images

AbstractThe automated detection of coasts, riverbanks, and polynyas from synthetic aperture radar images is a difficult image processing task due to speckle noise. In this work we present a novel Fuzzy-Wavelet framework for bordeline region detection in SAR images. Our technique is based on a combination of Wavelet denoising and Fuzzy Logic which boost decision-making on noisy and poorly defined environments. Unlike most recent filtering-detection algorithms, we do not apply hypothesis tests (Wilcoxon-Mann Whitney-G0) to label the edge point candidates one by one, instead we construct a fuzzy map from wavelet denoised image and extract their borderline. We compare our algorithm performance with the popular Frost–Sobel approach and a version of Canny’s algorithm with data-dependent parameters, over a database of real polynyas and coastline simulated images under the multiplicative model. The experimental results are evaluated by comparing Pratt’s Figure of Merit index of edge map quality. In almost all tes...

CAD-based 3D objects recognition in monocular images for mobile augmented reality

In order to improve registration performance of mobile augmented reality initialization, a new CAD-based recognition method of 3D objects in monocular images is proposed in this paper. Instead of estimating camera pose directly, our method attempts to estimate 3D object pose in the camera coordinate system. The geometric correspondence between the object mass center and its projection on 2D camera image is used to estimate object position. To hypothesize possible azimuths of the object, the 3D CAD model is off-screen rendered with OpenGL ES at different azimuths under an assumed inclination constraint provided by the inertial sensors of mobile devices, and compared with input image by contours matching. Unlike most recognition methods based on machine learning and natural features detection, which require large databases of real images and heavy computational consumption, our method only needs the CAD model of the 3D object in offline stage, and is computationally suitable for mobile devices. This paper describes the details of the method and shows its effectiveness with experiments.

No-reference image blur index based on singular value curve

We describe a new no-reference blur index for still images based on a singular value curve (SVC). The algorithm is composed of two steps. First, the singular value decomposition is performed on the image to be blur-assessed. Then an image blur index is constructed from the singular value curve. Experimental results obtained on four simulated blur databases and on the Real Blur Image Database show that the proposed SVC algorithm achieves high correlation against human judgments when assessing the blur distortion of images.

MsLRR: a unified multiscale low-rank representation for image segmentation.

In this paper, we present an efficient multiscale low-rank representation for image segmentation. Our method begins with partitioning the input images into a set of superpixels, followed by seeking the optimal superpixel-pair affinity matrix, both of which are performed at multiple scales of the input images. Since low-level superpixel features are usually corrupted by image noise, we propose to infer the low-rank refined affinity matrix. The inference is guided by two observations on natural images. First, looking into a single image, local small-size image patterns tend to recur frequently within the same semantic region, but may not appear in semantically different regions. The internal image statistics are referred to as replication prior, and we quantitatively justified it on real image databases. Second, the affinity matrices at different scales should be consistently solved, which leads to the cross-scale consistency constraint. We formulate these two purposes with one unified formulation and develop an efficient optimization procedure. The proposed representation can be used for both unsupervised or supervised image segmentation tasks. Our experiments on public data sets demonstrate the presented method can substantially improve segmentation accuracy.

Brain Stroke Detection by Microwaves Using Prior Information from Clinical Databases

Microwave tomographic imaging is an inexpensive, noninvasive modality of media dielectric properties reconstruction which can be utilized as a screening method in clinical applications such as breast cancer and brain stroke detection. For breast cancer detection, the iterative algorithm of structural inversion with level sets provides well-defined boundaries and incorporates an intrinsic regularization, which permits to discover small lesions. However, in case of brain lesion, the inverse problem is much more difficult due to the skull, which causes low microwave penetration and highly noisy data. In addition, cerebral liquid has dielectric properties similar to those of blood, which makes the inversion more complicated. Nevertheless, the contrast in the conductivity and permittivity values in this situation is significant due to blood high dielectric values compared to those of surrounding grey and white matter tissues. We show that using brain MRI images as prior information about brain's configuration, along with known brain dielectric properties, and the intrinsic regularization by structural inversion, allows successful and rapid stroke detection even in difficult cases. The method has been applied to 2D slices created from a database of 3D real MRI phantom images to effectively detect lesions larger than 2.5 × 10−2 m diameter.

Clustering via geometric median shift over Riemannian manifolds

The mean shift algorithms have been successfully applied to many areas, such as data clustering, feature analysis, and image segmentation. However, they still have two limitations. One is that they are ineffective in clustering data with low dimensional manifolds because of the use of the Euclidean distance for calculating distances. The other is that they sometimes produce poor results for data clustering and image segmentation. This is because a mean may not be a point in a data set. In order to overcome the two limitations, we propose a novel approach for the median shift over Riemannian manifolds that uses the geometric median and geodesic distances. Unlike the mean, the geometric median of a data set is one of points in the set. Compared to the Euclidean distance, the geodesic distances can better describe data points distributed on Riemannian manifolds. Based on these two facts, we first present a novel density function that characterizes points on a manifold with the geodesic distance. The shift of the geometric median over the Riemannian manifold is derived from maximizing this density function. After this, we present an algorithm for geometric median shift over Riemannian manifolds, together with theoretical proofs of its correctness. Extensive experiments have demonstrated that our method outperforms the state-of-the-art algorithms in data clustering, image segmentation, and noise filtering on both synthetic data sets and real image databases.

An optimized approach to histogram computation on GPU

A histogram is a compact representation of the distribution of data in an image with a full range of applications in diverse fields. Histogram generation is an inherently sequential operation where every pixel votes in a reduced set of bins. This makes finding efficient parallel implementations very desirable but challenging, because on graphics processing units thousands of threads may be atomically updating a short number of histogram bins. Under these circumstances, collisions among threads will be very frequent and such collisions will serialize thread execution, seriously damaging the performance. In this paper we propose a highly optimized approach to histogram calculation, which tackles such performance bottlenecks. It uses histogram replication for eliminating position conflicts, padding to reduce bank conflicts, and an improved access to input data called interleaved read access. Our so-called $${\mathcal{R}}$$ -per-block approach to histogram calculation has been successfully compared to the main state-of-the-art works using four histogram-based image processing kernels and two real image databases. Results show that our proposal is between 1.4 and 15.7 faster than every previous implementation for histograms of up to 4,096 bins.

Semi-supervised kernel learning based optical image recognition

This paper is to propose semi-supervised kernel learning based optical image recognition, called Semi-supervised Graph-based Global and Local Preserving Projection (SGGLPP) through integrating graph construction with the specific DR process into one unified framework. SGGLPP preserves not only the positive and negative constraints but also the local and global structure of the data in the low dimensional space. In SGGLPP, the intrinsic and cost graphs are constructed using the positive and negative constraints from side-information and k nearest neighbor criterion from unlabeled samples. Moreover, kernel trick is applied to extend SGGLPP called KSGGLPP by on the performance of nonlinear feature extraction. Experiments are implemented on UCI database and two real image databases to testify the feasibility and performance of the proposed algorithm.

Locality-Sensitive Hashing for Chi2 Distance

In the past 10 years, new powerful algorithms based on efficient data structures have been proposed to solve the problem of Nearest Neighbors search (or Approximate Nearest Neighbors search). If the Euclidean Locality Sensitive Hashing algorithm, which provides approximate nearest neighbors in a euclidean space with sublinear complexity, is probably the most popular, the euclidean metric does not always provide as accurate and as relevant results when considering similarity measure as the Earth-Mover Distance and 2 distances. In this paper, we present a new LSH scheme adapted to 2 distance for approximate nearest neighbors search in high-dimensional spaces. We define the specific hashing functions, we prove their local-sensitivity, and compare, through experiments, our method with the Euclidean Locality Sensitive Hashing algorithm in the context of image retrieval on real image databases. The results prove the relevance of such a new LSH scheme either providing far better accuracy in the context of image retrieval than euclidean scheme for an equivalent speed, or providing an equivalent accuracy but with a high gain in terms of processing speed.

An Image Processing Approach to Detect Lanes, Pot Holes and Recognize Road Signs in Indian Roads

Most of the Indian rural and sub urban roads are not ideal for driving due to faded lanes, irregular potholes, improper and invisible road signs. This has led to many accidents causing loss of lives and severe damage to vehicles. Many techniques have been proposed in the past to detect these problems using image processing methods. But there has been little work specifically carried out for detecting such issues of Indian roads. To address this acute problem, the study is undertaken with the objectives like, to make a survey of Indian roads, to suggest the method to detect lanes, potholes and road signs and their classification and to suggest automated driver guidance mechanism. In this regard, Hough Transformation method is adopted for Lane detection, where as Color Segmentation and Shape Modeling with Thin Spline Transformation (TPS) is used with nearest neighbor classifier for road sign detection and Classification. Further, K-means clustering based algorithm is adopted for pothole detection. Therefore, the attempt is made to invent an automated driver guidance mechanism to make the driving safe and easier in Indian roads. The experimental results obtained are tested with real time image database collected across different roads in sub-urban areas in India and found satisfactory.

Automated visual inspection of imprint quality of pharmaceutical tablets

Visual appearance is an important quality factor of pharmaceutical tablets. Moreover, it plays a key role in identification of tablets, which is needed to prevent mix-ups among various types of tablets. Since identification of tablets is most frequently done by imprints, good imprint quality, a property that makes the imprint readable, is of utmost importance in preventing mix-ups among the tablets. In this paper, we propose a novel method for automated visual inspection of tablets. Besides defect detection, imprint quality inspection is also considered. Performance of the method was evaluated on three different real tablet image databases of imprinted tablets. A “gold standard” was established by manually classifying tablets into a good and a defective class. The receiver operating characteristics (ROC) analysis indicated that the proposed method yields better sensitivity and specificity than the previous defect detection method.

No-reference image blur assessment using multiscale gradient

The increasing number of demanding consumer video applications, as exemplified by cell phone and other low-cost digital cameras, has boosted interest in no-reference objective image and video quality assessment (QA) algorithms. In this paper, we focus on no-reference image and video blur assessment. We consider natural scenes statistics models combined with multi-resolution decomposition methods to extract reliable features for QA. The algorithm is composed of three steps. First, a probabilistic support vector machine (SVM) is applied as a rough image quality evaluator. Then the detail image is used to refine the blur measurements. Finally, the blur information is pooled to predict the blur quality of images. The algorithm is tested on the LIVE Image Quality Database and the Real Blur Image Database; the results show that the algorithm has high correlation with human judgments when assessing blur distortion of images.

Transformation Invariant On-Line Target Recognition

Transformation invariant automatic target recognition (ATR) has been an active research area due to its widespread applications in defense, robotics, medical imaging and geographic scene analysis. The primary goal for this paper is to obtain an on-line ATR system for targets in presence of image transformations, such as rotation, translation, scale and occlusion as well as resolution changes. We investigate biologically inspired adaptive critic design (ACD) neural network (NN) models for on-line learning of such transformations. We further exploit reinforcement learning (RL) in ACD framework to obtain transformation invariant ATR. We exploit two ACD designs, such as heuristic dynamic programming (HDP) and dual heuristic dynamic programming (DHP) to obtain transformation invariant ATR. We obtain extensive statistical evaluations of proposed on-line ATR networks using both simulated image transformations and real benchmark facial image database, UMIST, with pose variations. Our simulations show promising results for learning transformations in simulated images and authenticating out-of plane rotated face images. Comparing the two on-line ATR designs, HDP outperforms DHP in learning capability and robustness and is more tolerant to noise. The computational time involved in HDP is also less than that of DHP. On the other hand, DHP achieves a 100% success rate more frequently than HDP for individual targets, and the residual critic error in DHP is generally smaller than that of HDP. Mathematical analyses of both our RL-based on-line ATR designs are also obtained to provide a sufficient condition for asymptotic convergence in a statistical average sense.

No-Reference Blur Assessment of Digital Pictures Based on Multifeature Classifiers

In this paper, we address the problem of no-reference quality assessment for digital pictures corrupted with blur. We start with the generation of a large real image database containing pictures taken by human users in a variety of situations, and the conduction of subjective tests to generate the ground truth associated to those images. Based upon this ground truth, we select a number of high quality pictures and artificially degrade them with different intensities of simulated blur (gaussian and linear motion), totalling 6000 simulated blur images. We extensively evaluate the performance of state-of-the-art strategies for no-reference blur quantification in different blurring scenarios, and propose a paradigm for blur evaluation in which an effective method is pursued by combining several metrics and low-level image features. We test this paradigm by designing a no-reference quality assessment algorithm for blurred images which combines different metrics in a classifier based upon a neural network structure. Experimental results show that this leads to an improved performance that better reflects the images' ground truth. Finally, based upon the real image database, we show that the proposed method also outperforms other algorithms and metrics in realistic blur scenarios.

No-reference blur index using blur comparisons

Presented is a new no-reference blur index for still images that is based on the observation that it can be difficult to perceive between versions of an image blurred to different degrees. A ‘re-blurred’ image is produced by intentionally blurring the test image. Local sample statistics are computed in the vicinity of detected edges of the original and re-blurred images, respectively. These are differenced and normalised to construct a new blur index. Experimental results on four simulated blur databases and on the Real Blur Image Database show that the proposed method obtains high correlations with test subjective quality evaluations.