Piotr M. Szczypinski

MaZda-A software package for image texture analysis

MaZda, a software package for 2D and 3D image texture analysis is presented. It provides a complete path for quantitative analysis of image textures, including computation of texture features, procedures for feature selection and extraction, algorithms for data classification, various data visualization and image segmentation tools. Initially, MaZda was aimed at analysis of magnetic resonance image textures. However, it revealed its effectiveness in analysis of other types of textured images, including X-ray and camera images. The software was utilized by numerous researchers in diverse applications. It was proven to be an efficient and reliable tool for quantitative image analysis, even in more accurate and objective medical diagnosis. MaZda was also successfully used in food industry to assess food product quality. MaZda can be downloaded for public use from the Institute of Electronics, Technical University of Lodz webpage.

Prediction of sensory texture quality attributes of cooked potatoes by NMR-imaging (MRI) of raw potatoes in combination with different image analysis methods

Nuclear magnetic resonance imaging (NMR-imaging), the so-called magnetic resonance imaging (MR-imaging), was performed on five potato varieties stored at 4 °C and 95% relative humidity for two and eight months, respectively. An image analysis on the obtained data and subsequent sensory analysis of the cooked potatoes displayed the high potential of employing advanced image analysis on MR-imaging data from raw potatoes to predict sensory attributes related to the texture of cooked potatoes. In contrast MR-imaging data were not found to correlate with specific gravity of the potatoes even though this parameter is normally found to correlate with the sensory texture quality of cooked potatoes. We suggest that this imply that MR-imaging beside giving well-known information about water distribution also gives information about anatomic structures within raw potatoes, which are of importance for the perceived textural properties of the cooked potatoes.

Texture and color based image segmentation and pathology detection in capsule endoscopy videos

This paper presents an in-depth study of several approaches to exploratory analysis of wireless capsule endoscopy images (WCE). It is demonstrated that versatile texture and color based descriptors of image regions corresponding to various anomalies of the gastrointestinal tract allows their accurate detection of pathologies in a sequence of WCE frames. Moreover, through classification of single pixels described by texture features of their neighborhood, the images can be segmented into homogeneous areas well matched to the image content. For both, detection and segmentation tasks the same procedure is applied which consists of features calculation, relevant feature subset selection and classification stages. This general three-stage framework is realized using various recognition strategies. In particular, the performance of the developed Vector Supported Convex Hull classification algorithm is compared against Support Vector Machines run in configuration with two different feature selection methods.

A model of deformable rings for interpretation of wireless capsule endoscopic videos

Wireless Capsule Endoscopy (WCE) provides a means to obtain a detailed video of the small intestine. A single session with WCE may produce nearly 8h of video. Its interpretation is tedious task, which requires considerable expertise and is very stressful. The Model of Deformable Rings (MDR) was developed to preprocess WCE video and aid clinicians with its interpretation. The MDR uses a simplified model of a capsules motion to flexibly match (register) consecutive video frames. Essentially, it computes motion-descriptive characteristics and produces a two-dimensional representation of the gastrointestinal (GI) tracts internal surface - a map. The motion-descriptive characteristics are used to indicate video fragments which exhibit segmentary contractions, peristalsis, refraction phases and areas of capsule retention. Within maps, certain characteristics that indicate areas of bleeding, ulceration and obscuring froth could be recognized. Therefore, the maps allow quick identification of such abnormal areas. The experimental results demonstrate that the number of discovered pathologies and gastrointestinal landmarks increases with the MDR technique.

Mazda - a software for texture analysis

This paper presents MaZda software for quantitative image texture analysis. This software, primarily developed for classification of magnetic resonance images, can be applied for wide class of textured images including color ones and 3D data. It enables estimation of almost 300 texture features; includes procedures for their reduction and classification. Feature clustering is also provided. The software has been developed since 1998. Currently it is a reliable and efficient tool used by many research institutes for different image analysis tasks.

Identifying barley varieties by computer vision

Abstract Visual discrimination between barley varieties is difficult, and it requires training and experience. The development of automatic methods based on computer vision could have positive implications for the food processing industry. In the brewing industry, varietal uniformity is crucial for the production of high quality malt. The varietal purity of thousands of tons of grain has to be inspected upon purchase in the malt house. This paper evaluates the effectiveness of identification of barley varieties based on image-derived shape, color and texture attributes of individual kernels. Varieties can be determined by means of discriminant analysis, including reduction of feature space dimensionality, linear classifier ensembles and artificial neural networks, with high balanced accuracy ranging from 67% to 86%. The study demonstrated that classification results can be significantly improved by standardizing individual kernel images in terms of their anteroposterior and dorsoventral orientation and performing additional analyses of wrinkled regions.

Quantitative analysis of lumbar intervertebral disc abnormalities at 3.0 Tesla: value of T2 texture features and geometric parameters

T2 relaxation time mapping provides information about the biochemical status of intervertebral discs. The present study aimed to determine whether texture features extracted from T2 maps or geometric parameters are sensitive to the presence of abnormalities at the posterior aspect of lumbar intervertebral discs, i.e. bulging and herniation. Thirty‐one patients (21 women and 10 men; age range 18–51 years) with low back pain were enrolled. MRI of the lumbar spine at 3.0 Tesla included morphological T1‐ and T2‐weighted fast spin‐echo sequences, and multi‐echo spin‐echo sequences that were used to construct T2 maps. On morphological MRI, discs were visually graded into ‘normal’, ‘bulging’ or ‘herniation’. On T2 maps, texture analysis (based on the co‐occurrence matrix and wavelet transform) and geometry analysis of the discs were performed. The three T2 texture features and geometric parameters best‐suited for distinguishing between normal discs and discs with bulging or herniation were determined using Fisher coefficients. Statistical analysis comprised ANCOVA and post hoc t‐tests. Eighty‐two discs were classified as ‘normal’, 49 as ‘bulging’ and 20 showed ‘herniation.’ The T2 texture features Entropy and Difference Variance, and all three pre‐selected geometric parameters differed significantly between normal and bulging, normal and herniated, and bulging and herniated discs (p < 0.05). These findings suggest that T2 texture features and geometric parameters are sensitive to the presence of abnormalities at the posterior aspect of lumbar intervertebral discs, and may thus be useful as quantitative biomarkers that predict disease. Copyright

Computer Simulation of Magnetic Resonance Angiography Imaging: Model Description and Validation

With the development of medical imaging modalities and image processing algorithms, there arises a need for methods of their comprehensive quantitative evaluation. In particular, this concerns the algorithms for vessel tracking and segmentation in magnetic resonance angiography images. The problem can be approached by using synthetic images, where true geometry of vessels is known. This paper presents a framework for computer modeling of MRA imaging and the results of its validation. A new model incorporates blood flow simulation within MR signal computation kernel. The proposed solution is unique, especially with respect to the interface between flow and image formation processes. Furthermore it utilizes the concept of particle tracing. The particles reflect the flow of fluid they are immersed in and they are assigned magnetization vectors with temporal evolution controlled by MR physics. Such an approach ensures flexibility as the designed simulator is able to reconstruct flow profiles of any type. The proposed model is validated in a series of experiments with physical and digital flow phantoms. The synthesized 3D images contain various features (including artifacts) characteristic for the time-of-flight protocol and exhibit remarkable correlation with the data acquired in a real MR scanner. The obtained results support the primary goal of the conducted research, i.e. establishing a reference technique for a quantified validation of MR angiography image processing algorithms.

Convex Hull-Based Feature Selection in Application to Classification of Wireless Capsule Endoscopic Images

In this paper we propose and examine a Vector Supported Convex Hull method for feature subset selection. Within feature subspaces, the method checks locations of vectors belonging to one class with respect to the convex hull of vectors belonging to the other class. Based on such analysis a coefficient is proposed for evaluation of subspace discrimination ability. The method allows for finding subspaces in which vectors of one class cluster and they are surrounded by vectors of the other class. The method is applied for selection of color and texture descriptors of capsule endoscope images. The study aims at finding a small set of descriptors for detection of pathological changes in the gastrointestinal tract. The results obtained by means of the Vector Supported Convex Hull are compared with results produced by a Support Vector Machine with the radial basis function kernel.

Selecting texture discriminative descriptors of capsule endpscopy images

In supervised data classification one of the problems is to reduce dimensionality of feature vectors. It is important to find such features which have high ability for discrimination of diverse classes and to get rid of features which are useless for such discrimination. In this paper we propose a new method for feature subset selection utilizing a convex hull (or convex polytope). The method searches for feature space subspaces in which vectors of one class are surrounded by vectors of the other class. The method is applied for selection of color and texture descriptors of capsule endoscope images. The study aims at finding a small set of descriptors for detection of pathological changes in the gastrointestinal tract. The results are compared with results produced by a Radial Basis Function Network classifier.