Run-length Histogram Research Articles

Texture analysis of sonographic image of placenta in pregnancies with normal and adverse outcomes, a pilot study

IntroductionMany maternal and fetal morbidity and mortality from complications of pregnancy have been attributed to placenta abnormality. Placenta assessment in developing countries is mainly through ultrasonography which is subjective and prone to error. Objective assessment of placental abnormalities through texture analysis has been frequently done using magnetic resonance images with little done on ultrasound generated images, thus, the need for this study. The study is aimed at using statistical texture analysis in characterizing placenta tissue into normal and abnormal placenta as well as testing the accuracy of different texture analysis algorithms in differentiating placenta into normal and abnormal placental tissues. MethodsThis longitudinal study involved 500 ultrasound-generated placenta images from patients screened for adverse pregnancy outcomes in a private hospital in Enugu. These images were loaded onto an HP laptop for viewing. Two regions of interest were selected from the placenta tissue where texture features were extracted and were classified into normal and abnormal placentas using MaZda® software version 47 while the accuracy of the classification descriptors was assessed using WEKA classification algorithms. ResultsCo-occurrence matrix, run length matrix and histogram parameters differentiated normal placenta tissue from abnormal placental tissues (p-value <0.05) while variance is the only absolute gradient parameter that can differentiate normal placenta tissue from abnormal placenta tissues. All feature descriptors show high classification accuracy using KNN and ANN algorithms. ConclusionTexture analysis can differentiate normal placenta tissues from abnormal placenta tissue which will reduce the errors associated with subjective assessment of the placenta echogenicity. Implications for practiceIntegrating these computer-aided algorithms into our ultrasound machines will lead to early detection of abnormal placenta tissues as early management results in better pregnancy outcomes.

A texture-based approach for word script and nature identification

In this work, we propose a texture-based approach to separate handwritten from machine-printed words, written in Arabic and Latin scripts. The idea is to benefit from differences in writing orientation and the difference between the stroke length to discriminate between these scripts. For that, we designed a K nearest neighbors classifier trained with a set of texture features. These features are extracted from black run-length (BRL) histograms and seem to be suitable for finding structural characteristics in word images. Four feature extraction scenarios: (1) BRL, (2) restricted BRL, (3) BRL statistics and (4) restricted BRL combined to their statistics are chosen to demonstrate the potential of such a texture-based approach in script identification. Exploiting these features, we have got very promising result. The identification correct rate is higher than 98.92 % in our experiments.

Contrast Modification Forensics Algorithm Based on Merged Weight Histogram of Run Length

A novel image forensic algorithm against contrast modification based on merged weight histogram of run length is proposed. In the proposed algorithm, the run length histogram features were firstly extracted, and then those of different orientation were subsequently merged; after normalization of the prior features, the authors calculated leaps in the histogram numerically; lastly, the generated features of authentic and tampered images were trained by a SVM classifier. Large amounts of experiments show that, the proposed algorithm has low cost of computation complexity, compared with some existing scheme, and it has better performance with many test databases, furthermore, the proposed algorithm can effectively detect local contrast modification of image.

SU‐E‐J‐259; How Does CT Reconstruction Kernel Affect the Radiogenomic Features in Non‐Small Cell Lung Cancer?

Purpose:Radiogenomics promises to discover quantitative imaging features which are associated with genomic profiles and of prognostic in cancer patients. However, the CT imaging features are known to be sensitive to noise characteristic and affected by CT parameters. We investigate the variability of CT imaging features which are previously reported as radiogenomic markers in non‐small cell lung cancer (NSCLC).Methods:Three NSCLC cases of CT exams with 2 reconstruction kernels (B30f, B60f) were selected. We set ten ROIs within tumor mask. Run length nonuniformity (RLN) and histogram energy, which were previously reported to be predictive of patient survival, were calculated for each ROI. For establishing the intrinsic variability of image feature depending on the reconstruction kernel, we carried out the same feature extraction procedure on the COPDGene phantom CT images. The coefficient of variation (CoV) was compared between imaging features from images of single reconstruction kernel and mixed kernel image set.Results:RLN and Energy of standard kernel for phantom data showed higher values up to 366% than that of sharp kernel. For patient cases, the image features of standard kernel was higher 196% than that of sharp kernel on an average. The CoV of RLN was 0.142±0.035 for B30f data set and 0.060±0.015 for B60f data set in patient case study. For the mixed kernel dataset, the CoV of RLN increased to 0.407±0.061. The CoV of Energy feature values showed similar trends with RLN.Conclusion:Out study revealed high variability of CT image features depending on reconstruction kernel. These characteristic should be considered in feature extraction procedure when different kernels are used in the patient dataset. Use of the same CT protocol is preferred. Otherwise, application of kernel normalization techniques is necessary in the radiogenomic study.

A statistical-textural-features based approach for classification of solid drugs using surface microscopic images.

The quality of pharmaceutical products plays an important role in pharmaceutical industry as well as in our lives. Usage of defective tablets can be harmful for patients. In this research we proposed a nondestructive method to identify defective and nondefective tablets using their surface morphology. Three different environmental factors temperature, humidity and moisture are analyzed to evaluate the performance of the proposed method. Multiple textural features are extracted from the surface of the defective and nondefective tablets. These textural features are gray level cooccurrence matrix, run length matrix, histogram, autoregressive model and HAAR wavelet. Total textural features extracted from images are 281. We performed an analysis on all those 281, top 15, and top 2 features. Top 15 features are extracted using three different feature reduction techniques: chi-square, gain ratio and relief-F. In this research we have used three different classifiers: support vector machine, K-nearest neighbors and naïve Bayes to calculate the accuracies against proposed method using two experiments, that is, leave-one-out cross-validation technique and train test models. We tested each classifier against all selected features and then performed the comparison of their results. The experimental work resulted in that in most of the cases SVM performed better than the other two classifiers.

Large-scale document image retrieval and classification with runlength histograms and binary embeddings

We present a new document image descriptor based on multi-scale runlength histograms. This descriptor does not rely on layout analysis and can be computed efficiently. We show how this descriptor can achieve state-of-the-art results on two very different public datasets in classification and retrieval tasks. Moreover, we show how we can compress and binarize these descriptors to make them suitable for large-scale applications. We can achieve state-of-the-art results in classification using binary descriptors of as few as 16–64 bits.

A high-capacity data-hiding approach for polygonal meshes using maximum expected level tree

This paper presents a high-capacity data-hiding approach for polygonal meshes using maximum expected level tree based on a message probability model. While embedding/extracting a bitstring for a primitive, a maximum expected level tree is built level-by-level for all remaining primitives arranged in a reference order. At each level, the number of leaf nodes in one subtree is determined based on the occurrence probability of the interested event for the next message bit from the proposed message probability model, such that the traversal path from the root to the embedding/extracting primitive (leaf node) is as long as possible. Moreover, the traversal path represents the embedded/extracted message. Given an embedding message, the 0-bit and 1-bit run-length histograms on which the proposed message probability model is based are computed initially and updated after each embedding/extracting step. When extracting, the histograms are extracted first from the stego model and then used to extract the message from each primitive.The capacity of our approach depends on the run-length histograms of a given embedding message. The experimental results show that our capacity is larger than that of previous methods, the improvements in capacity for text-based, binary-based, and compressed-based message types are approximately 18.21%, 44.79%, and 14.86%, respectively. Moreover, as a permutation steganography approach, the stego model is distortion-free, and the embedded message is imperceptible and robust to geometric affine transformations.

Reversible binary image data hiding by histogram modification

A novel reversible binary image data hiding scheme using Run-Length(RL) histogram modification was presented.Select the sequence of RL couples in which the length of couple is not shorter than threshold T1: First eliminate the white RL of isolated white pixels,and then embed data in the black RL whose length equals T.The proposed scheme worked for text,graphics,their mixture,and both halftone and non-halftone binary images.Experimental results have shown its superiority to other methods.

A Block-Edge-Pattern-Based Content Descriptor in DCT Domain

In this correspondence, we describe a robust and effective content descriptor based on block-edge patterns extracted in discrete cosine transform domain, which is suitable for applications in JPEG or MPEG compressed images and videos. This content descriptor is constructed by a run-length edge-block histogram with three patterns including horizontal edge, vertical edge and no edge. In comparison with existing descriptors, the proposed features: 1) low-cost computing suitable for real-time implementation and high-speed processing of compressed videos; 2) robust to orientation changes such as rotation, noise, reverse, etc.; 3) operates in compressed domain. Extensive experiments support that the proposed content descriptor is effective in describing visual content, and achieves superior performances in terms of retrieval precision and recall rates.

Color text image binarization based on binary texture analysis

In this paper, a novel binarization algorithm for color text images is presented. This algorithm effectively integrates color clustering and binary texture analysis, and is capable of handling situations with complex backgrounds. In this algorithm, dimensionality reduction and graph theoretical clustering are first employed. As the result, binary images related to clusters can be obtained. Binary texture analysis is then performed on each candidate binary image. Two kinds of effective texture features, run-length histogram and spatial-size distribution related respectively, are extracted and explored. Cooperating with an LDA classifier, the optimal candidate of the best binarization effect is obtained. Experiments with images collected from Internet has been carried out and compared with existing techniques, both show the effectiveness of the algorithm.

Color text image binarization based on binary texture analysis

In this paper, a novel binarization algorithm for color text images is presented. This algorithm effectively integrates color clustering and binary texture analysis, and is capable of handling situations with complex backgrounds. In this algorithm, dimensionality reduction and graph theoretical clustering are first employed. As the result, binary images related to clusters can be obtained. Binary texture analysis is then performed on each candidate binary image. Two kinds of effective texture features, run-length histogram and spatial-size distribution related respectively, are extracted and explored. Cooperating with an LDA classifier, the optimal candidate of the best binarization effect is obtained. Experiments with images collected from Internet has been carried out and compared with existing techniques, both show the effectiveness of the algorithm.

An adaptive logical method for binarization of degraded document images

This paper describes a modified logical thresholding method for binarization of seriously degraded and very poor quality gray-scale document images. This method can deal with complex signal-dependent noise, variable background intensity caused by nonuniform illumination, shadow, smear or smudge and very low contrast. The output image has no obvious loss of useful information. Firstly, we analyse the clustering and connection characteristics of the character stroke from the run-length histogram for selected image regions and various inhomogeneous gray-scale backgrounds. Then, we propose a modified logical thresholding method to extract the binary image adaptively from the degraded gray-scale document image with complex and inhomogeneous background. It can adjust the size of the local area and logical thresholding level adaptively according to the local run-length histogram and the local gray-scale inhomogeneity. Our method can threshold various poor quality gray-scale document images automatically without need of any prior knowledge of the document image and manual fine-tuning of parameters. It keeps useful information more accurately without overconnected and broken strokes of the characters, and thus, has a wider range of applications compared with other methods.

Document image binarization based on texture features

Binarization has been difficult for document images with poor contrast, strong noise, complex patterns, and/or variable modalities in gray-scale histograms. We developed a texture feature based thresholding algorithm to address this problem. Our algorithm consists of three steps: 1) candidate thresholds are produced through iterative use of Otsu's algorithm (1978); 2) texture features associated with each candidate threshold are extracted from the run-length histogram of the accordingly binarized image; 3) the optimal threshold is selected so that desirable document texture features are preserved. Experiments with 9,000 machine printed address blocks from an unconstrained US mail stream demonstrated that over 99.6 percent of the images were successfully binarized by the new thresholding method, appreciably better than those obtained by typical existing thresholding techniques. Also, a system run with 500 troublesome mail address blocks showed that an 8.1 percent higher character recognition rate was achieved with our algorithm as compared with Otsu's algorithm.

Single cytoplasmic dynein molecule movements: characterization and comparison with kinesin

Cytoplasmic dynein is a major microtubule motor for minus-end directed movements including retrograde axonal transport. To better understand the mechanism by which cytoplasmic dynein converts ATP energy into motility, we have analyzed the nanometer-level displacements of latex beads coated with low numbers of cytoplasmic dynein molecules. Cytoplasmic dynein-coated beads exhibited greater lateral movements among microtubule protofilaments (ave. 5.1 times/microns of displacement) compared with kinesin (ave. 0.9 times/micron). In addition, dynein moved rearward up to 100 nm over several hundred milliseconds, often in correlation with off-axis movements from one protofilament to another. We suggest that single molecules of cytoplasmic dynein move the beads because 1) there is a linear dependence of bead motility on dynein/bead ratio, 2) the binding of beads to microtubules studied by laser tweezers is best fit by a first-order Poisson, and 3) the run length histogram of dynein beads follows a first-order decay. At the cellular level, the greater disorder of cytoplasmic dynein movements may facilitate transport by decreasing the duration of collisions between kinesin and cytoplasmic dynein-powered vesicles.

Understanding the process of quantitative ultrasonic tissue characterization.

Because the human vision system cannot distinguish the broad range of gray values that a computer visual system can, computerized image analysis may be used to obtain quantitative information from ultrasonographic (US) real-time B-mode scans. Most quantitative US involves programming an off-line computer to accept, analyze, and display US image data in a way that enhances the detection of changes in small-scale structures and blood flow that occur with disease. Common image textural features used in quantitative US tissue characterization consist of first-order gray-level statistics (eg, occurrence frequency of gray levels independent of location or spatial relationship) and second-order gray-level statistics dependent on location and spatial relationship, including statistical analysis of gradient distribution, co-occurrence matrix, covariance matrix, run-length histogram, and fractal features. A customized tissue signature software has been developed to analyze image data obtained from clinical US scanners. Means comparison testing and multivariate analysis techniques are used to compare the numbers generated for a particular region of interest. By integrating these techniques into the radiologist's interpretation of the sonogram, the quantitative information gained may lead to earlier detection of lesions difficult to see with the human eye.

IX. MR tissue characterization of intracranial tumors by means of texture analysis

Tissue characterization of the human brain has been performed by texture analysis of proton relaxation time images using a standard MR whole body imager operating at 1.5 T. A combined CP CPMG multiecho, multislice sequence was used to measure T 1 and T 2 in each pixel with an uncertainty not exceeding 10%. In a prospective clinical study, 12 patients with histologically confirmed brain tumors were investigated. For each ROI in the calculated T 1 and T 2 parameter images, texture parameters originating from the grey level distribution, the gradient distribution, the grey level co-occurrence matrix, and the grey level runlength histogram were used for classification and discrimination between tissues. All regions corresponding to the normal brain tissue (white matter, grey matter, cerebrospinal fluid) were successfully discriminated from each other as well as from the pathological tissue parts (edema and tumor). The classification of 10 edematous and 8 tumorous tissue regions yielded only one misclassification. Together with additional rules, these discrimination rules formed the knowledge base of an expert system for segmentation of the brain images. In cases of tumors without Gd-DTPA contrast medium uptake or in cases of Gd-DTPA contraindication, segmentated images can help solve nontrivial diagnostical problems such as delineating the target volume in radiation therapy.

Magnetic Resonance Imaging of Intracranial Tumors: Tissue Characterization by Means of Texture Analysis

Tissue characterization of the human brain has been performed by texture analysis of proton relaxation time images using a standard MRI whole body imager operating at 1.5 T. A combined CP/CPMG multiecho, multislice sequence was used to measure T1 and T2 in each pixel with an uncertainty not exceeding 10 %. In a prospective clinical study twelve patients with histologically confirmed brain tumors were investigated. For each ROI in the calculated T1 and T2 parameter images, texture parameters originating from the greylevel distribution, the gradient distribution, the greylevel co-occurence matrix and the greylevel runlength histogram were used for classification and discrimination between tissues. All regions corresponding to the normal brain tissues (white matter, grey matter, cerebrospinal fluid) were successfully discriminated from each other as well as from the pathological tissue parts (edema and tumor). The classification of ten edematous and eight tumorous tissue regions yielded only one misclassification. Together with additional rules these discrimination rules formed the knowledge base of an expert system for segmentation of the brain images. In cases of tumors without Gd-DTPA contrast medium uptake or in cases of Gd-DTPA contraindication, segmented images can support to solve nontrivial diagnostic problems such as to delineate the target volume in radiation therapy.

Computer-aided evaluation of ultrasonic images of the thyroid

Forty-two patients are described. The ultrasound patterns of normal and pathologically changed thyroid glands have been analysed with the aid of a computer. Using several parameters such as run-length histograms, brightness-histograms, middle run length and standard deviation, we are able to classify the patterns of a thyroid lobe or of a part of it into 4 or 5 major groups. The estimation of the benign or malignant character of the lesions with the used parameters must be controlled by further investigation.

Ancient Hebraic Handwriting Identification with Run-Length Histograms

Ancient Hebraic Handwriting Identification with Run-Length Histograms

Distribution of run lengths over scanned rectangles

Two of the most common operations performed on pictorial data are thresholding to delineate areas of interest, and run-length encoding of similar regions along a scan line in a television-like scan of this binary data. This run-length encoding of binary images requires little auxiliary storage and can often be performed almost simultaneously with the scanning operation. In this paper, we show that the information-compressive run-length histogram, commonly used to set up the code tables for Huffman encoding, carries significant shape information for a population of scanned rectangles. Our method is to decompose the problem by introducing auxiliary variables and finding conditional probability densities of run lengths using these variables. Probability distributions are found for run lengths derived from a line-by-line scan of a population of uniform intensity rectangles of one size but arbitrary orientation on a background of a different uniform intensity. For this idealized situation, the probability densities have two singularities. New, more robust estimates of rectangle shape derived from the run-length histogram are given. Sources of these singularities are examined and used to extend the results to more complex and less idealized situations. We conclude that the probability density of run lengths and similar random variables over similar populations will still have peaks or singularities. Thus sampled representations such as histograms may be very sensitive to sampling methods and parameters.