Search

Background: Direct detection of polyps from colon capsule endoscopy (CCE) videos is an ultimate goal not only for physicians but also for biomedical engineers who are working on automatic internal lesions like polyps. There is also a great enthusiasm among biomedical professionals to make advanced systems for aiding doctors to have a faster and accurate diagnosis by the means of polyp detection from CCE acquired video streams. Such systems must be able to localize polyps correctly and extract the whole lesions from the video frames completely.Material and Methods: In this paper, a new approach toward object-wise polyp detection from CCE frames in a video stream is proposed. The proposed method employs modified region proposal CNNs to localize the existing polyps from CCE acquired video frames and after that a level-set method known as Distance Regularized Level Set Evolution (DRLSE) is employed for automatic model-based segmentation of localized polyps. The pixel-wise detection of polyps is necessary for polyp classification and will help gastroenterologists to determine appropriate prognosis and treatment for the patients.Results and conclusion: The proposed method is trained by the means of an CCE still image dataset which includes manually annotated polyps. The trained network is then applied to CCE video images. The results demonstrate that the proposed method is able to localize and detect polyps both region-wise and pixel-wise with a good rate of accuracy.

A 3D ultrasound computer tomography (USCT) device with a nearly isotropic and spatially invariant 3D point spread function has been constructed at Institute for Data Processing and Electronic (IPE), Karlsruhe Institute of Technology (KIT). This device is currently applied in clinical studies for breast cancer screening. In this paper, a new method to develop an automated segmentation algorithm for USCT acquired images is proposed. The method employs distance regularized level set evolutionary (DRLSE) active contours along with surface fitting extrapolation and 3D binary mask generation for fully automatic segmentation outcome. In the first stage of the proposed algorithm, DRLSE is applied to those 3D USCT slice images which contain breast and are less affected by noise and ring artifacts named as Cat2. The DRLSE segmentation results are employed to extrapolate the rest of slice images known as Cat1. To overcome defectively segmented slice images, a 3D binary mask is generated out of USCT attenuation images. The 3D binary mask is multiplied by the DRLSE-based segmentation results to form finally segmented 3D USCT images. The method was tested on 12 clinical dataset images. According to F-measure criterion, the proposed method shows higher performance than the previously proposed semiautomatic segmentation one.

Display OmittedThe block diagram of the proposed system. Occasionally certain areas in the retina can be questionable for physicians which can lead to wrong interpretations for patients.A method is proposed that introduces a higher ability of segmentation by employing Skeletonization and a threshold selection based on Fuzzy Entropy.By extracting indices of the human retina properly, physicians will be able to estimate pathological injuries with a higher confidence.The proposed approach is fast and outperforms over other previously competitive techniques.The proposed approach consists of two stages. First of all, the retinal vessels was preprocessed by the HSV space and Wiener Filter. Then, the segmentation level is implemented by using Adaptive Filter that employs optimum threshold based on Fuzzy Entropy and Skeleton algorithm. The analysis of retina blood vessels in clinics indices is one of the most efficient methods employed for diagnosing diseases such as diabetes, hypertension and arthrosclerosis. In this paper, an efficient algorithm is proposed that introduces a higher ability of segmentation by employing Skeletonization and a threshold selection based on Fuzzy Entropy. In the first step, the blurring noises caused by hand shakings during ophthalmoscopy and color photography imageries are removed by a designed Wieners filter. Then, in the second step, a basic extraction of the blood vessels from the retina based on an adaptive filtering is obtained. At the last step of the proposed method, an optimal threshold for discriminating main vessels of the retina from other parts of the tissue is achieved by employing fuzzy entropy. Finally, an assessment procedure based on four different measurement techniques in the terms of retinal fundus colors is established and applied to DRIVE and STARE database images. Due to the evaluation comparative results, the proposed extraction of retina blood vessels enables specialists to determine the progression stage of potential diseases, more accurate and in real-time mode.

Colorectal lesions known as polyps are one of the diagnostic symptoms for colorectal disease. So, their accurate detection and localization based on a computer-aided diagnosis can assist colonists for prescribing more effective treatments. The computer vision and machine learning methods like pattern recognition and deep learning neural networks are the most popular strategies for automatic polyp detection purpose. The proposed approach in this paper is an innovative deep learning neural network. The proposed network has a novel U-Net architecture. The architecture of proposed network includes fully 3D layers which enable the network to be fed with multi or hyperspectral images or even video streams. Moreover, there is a dice prediction output layer. This type of output layer employs probabilistic approaches and benefits from more accurate prediction abilities. The proposed method is applied to international standard optical colonoscopy datasets known as CVC-ClinicDB, CVC-ColonDB and ETIS-Larib. The implementation and evaluation results demonstrate that the proposed U-Net outperforms other competitive methods for automatic polyp detection based on accuracy, precision, recall and F-Score criteria. The proposed method can assist experts and physicians to localize colonial polyps with more accuracy and speed. In addition, the proposed network can be used on live colonoscopy observations due to its high performance and fast operability.

In recent years, employment of dispatching systems including SCADA in Iranian Regional Electric Companies for automation of power plants, sites and generators has become common and outspread. Such systems comprise of some specific equipments such as remote terminal unit (RTU), communication medium and computerized dispatching center. For connecting RTU to computer center, different telecommunication media are employed. These media can be power line carrier (PLC), fiber optic and other conventional communication methods. In some cases, there is no way to employ the named communication methods due to unavailability of essential media. In such cases, it is necessary that a new communication strategy be established. In this paper, two new strategies for providing possible data transmission are proposed. For the first solution, a two-way communication over serial to Ethernet conversion base on static IP is proposed. In the first strategy, the communication is provided through asynchronous digital subscriber line (ADSL) modem. The second strategy is proposed for data exchanging between distributed generators (DGs) and upstream electric power stations. The acquired and collected data are transmitted to dispatching center. The proposed telecommunication strategies are implemented practically with success and acceptable performance for DGs under supervision of Fars regional electric company (FREC).

Providing stable and robust power signals for electrical consumers and apparatuses is the most important responsibility of all electric power providers. Whenever the electric power signals suffer from disturbances which affect their quality and consequently peril the safety and right operation of electrical appliances, it is the main task of suppliers to detect and solve such obstacles. For defect prevention and faulty situation treatment caused by power quality disturbances, it is necessary to detect and classifythem in a reliable and guaranteed manner. In this paper, an innovative approach toward confident classification of four distinct types of power quality disturbances is proposed. The proposed method comprises of two main stages. In the first stage, noise resistive and steady features based on a new one dimensional local binary pattern approach are extracted and the desired feature vectors are formed. The second stage devotes to the reliable classification of the feature vectors belonging to the studied power quality disturbances based on conventional neural networks. The evaluation results are implemented in the form of Precision, Recall and F-measure. The F-measure about 91% demonstrates the higher efficiencyand performance of proposed method in comparison to the previously proposed strategies based on discrete wavelet and some statistical features with the same neural network classification.

In modern remote sensing procedures, one of the most important issues is to distinguish specific types of land coverage. Discrimination between different land coverages especially in metropolitan surveying is so important that the in front civilization projects are basically dependent to them. In this paper, an innovative image processing strategy is employed for distinguishing green lands from other metropolitan areas in aerial imaging. The main purpose of this project is to audit green land areas, either public or private, for forthcoming municipal projects. The proposed method is constituted of four main stages. In the first step, the acquired aerial video frames, even in or offline modes, are converted into static images. In the second and third stages, two distinct pre-processing stages are deployed. The output of these two preprocessing stages is segmented into two parts comprising of green land and other urban areas. The evaluation and experimental results demonstrate the fair and applicable performance near 90% (86% in average) in F-score criterion.

This paper introduces a computer-assisted diagnosis (CAD) system for automatic mitosis detection from breast cancer histopathology slide images. In this system, a new approach for reducing the number of false positives is proposed based on Teaching-Learning-Based optimization (TLBO). The proposed CAD system is implemented on the histopathology slide images acquired by Aperio XT scanner (scanner A). In TLBO algorithm, the number of false positives (falsely detected nonmitosis candidates as mitosis ones) is defined as a cost function and, by minimizing it, many of nonmitosis candidates will be removed. Then some color and texture (textural) features such as those derived from cooccurrence and run-length matrices are extracted from the remaining candidates and finally mitotic cells are classified using a specific support vector machine (SVM) classifier. The simulation results have proven the claims about the high performance and efficiency of the proposed CAD system.

Histopathology slides are one of the most applicable resources for pathology studies. As observation of these kinds of slides even by skillful pathologists is a tedious and time-consuming activity, computerizing this procedure aids the experts to have faster analysis with more case studies per day. In this paper, an automatic mitosis detection system (AMDS) for breast cancer histopathological slide images is proposed. In the proposed AMDS, the general phases of an automatic image based analyzer are considered and in each phase, some special innovations are employed. In the pre-processing step to segment the input digital histopathology images more precisely, 2D anisotropic diffusion filters are applied to them. In the training segmentation phase, the histopathological slide images are segmented based on RGB contents of their pixels using maximum likelihood estimation. Then, the mitosis and non-mitosis candidates are processed and hence that their completed local binary patterns are extracted object-wise. For the classification phase, two subsequently non-linear support vector machine classifiers are trained pixel-wise and object-wise, respectively. For the evaluation of the proposed AMDS, some object and region based measures are employed. Having computed the evaluation criteria, our proposed method performs more efficient according to f-measure metric (70.94% for Aperio XT scanner images and 70.11% for Hamamatsu images) than the methods proposed by other participants at Mitos-ICPR2012 contest in breast cancer histopathological images. The experimental results show the higher performance of the proposed AMDS compared with other competitive systems proposed in Mitos-ICPR2012 contest.

 Abstract—Automatic and reliable extraction of the minutiae from fingerprint images is a critical process in fingerprint matching and a main preprocess for this stage is Thinning. There are a lot of algorithms for fingerprint thinning procedure. All of the previously proposed thinning methods try to thin every ridge due to the content of its central pixel and then extracting minutiae based on some other algorithms for denoising and preventing false minutiae detections at islands or spurities. If an algorithm could thin fingerprint ridges except unrecoverable corrupted regions and also could eliminate noise, it will be considered as a good thinning one and no additional processes are needed before minutiae extraction. The proposed method of this paper has such abilities. The proposed algorithm is implemented by applying four boxes of matrices; each of them thins ridges due to a specific direction; i.e., diagonal, horizontal and vertical directions. The proposed algorithm also is able to thin discrete Latin Characters or symbols. For evaluating the proposed method, several robust and reliable experiments have been employed and the results confirm the higher ability of the proposed method in comparison with the other competing one.

In this paper, a new approach for fingerprint ridge orientation field matching based on a novel HMM (Hidden Markov Model) is proposed. The proposed method comprises several steps. First steps are devoted to regular fingerprint preprocesses and ridge orientation estimation. Then, the fingerprint images are registered along a reference point. Next, the proposed HMM topology is applied to the predetermined fingerprint orientation field information around the reference point. The suggested HMM is of improved training abilities. After applying the proposed HMM to the ridge orientation field, the matching cells are produced. These cells consist of transition, observation and initial probability matrices which will be used in the matching procedure. The proposed matching method has been evaluated using some creditable fingerprint databases such as FVC2000 DB2_A, FVC2004 DB3_A and DB4_A. The evaluation results confirm higher efficiency, robustness and accuracy for the proposed method compared with the previously proposed matching ones.

In this paper, an innovative blind watermarking algorithm has been proposed for imagery applications. This algorithm has used the coefficients of the discrete wavelet transform of the host image in the form of super trees to embed the predefined binary watermark in the host image. In this scheme, a pseudo random sequence is generated to determine the exact wavelet super trees used for embedding procedure. In the next step, after choosing the maximum and second maximum amplitude coefficients of each super tree, the distance vector between two coefficients is computed. For embedding bit zero of the specified watermark, the values of the distance vector elements are decreased, while for embedding bit 1, those values will be increased based on the proposed formulas. The experimental results show that the proposed algorithm has significant robustness against image processing attacks, especially JPEG compression and also the PSNR value for the watermarked images generated by the proposed method is more than 42 dB. The watermark bits may be inserted into the host image either in spatial or in the transform domain. It is shown that hiding information in the transform domain will lead to a more robust watermarking system against most of the attacks (6). Among these methods, methods based on discrete wavelet transform (DWT) are of great renown. This is due to specific features of space-frequency localization, multi-resolution display, linear computational complexity and the excellent modeling related to Human Vision System

Nowadays, automatic computer-Aided Diagnosis (CAD) systems for grading different types of cancers like breast cancer are very prevalent. These systems employ histopathology slide images acquired by advanced and well-defined digital scanners. The previously proposed automatic or computer-aided systems for breast cancer grading, especially by counting mitoses, suffer from various types of deficiencies. The most important one is their low efficiency along with high complexity due to the huge amount of features. In this paper, two types of features with more flexibility and less complexity are employed. These features are Completed Local Binary Pattern (CLBP) as textural features and Stiffness Matrix as geometric, morphometric and shape-based features. In the proposed automatic mitosis detection system, these two features are fused with each other. The evaluation results are for histology Dataset H (Hamamatsu Nanozoomer Scanners) provided by Mitos-ICPR2012 contest sponsors. Employing a nonlinear RBF kernel support vector machine (SVM) classifier with parameter sigma which equals to 100, leads to an efficiency of 82%. The results are in the form of F-measure criterion which is a reliable and mostly common evaluation criterion for such biological systems.

This paper proposes a lossless data hiding (LDH) scheme on uncompressed video data based on a multi-level histogram shifting mechanism in integer wavelet transform (IWT) domain. The proposed method enables the exact recovery of the original host signal upon extracting the embedded information, if the watermarked image is not affected by any other process. In the proposed scheme, the approximation subband image of the luminance component of a video frame is computed. Then, it is divided into non overlapping blocks. In each block, the differences between the neighboring elements are computed and a histogram is made on the difference values. The secret data are embedded into the blocks based on a multi-level shifting mechanism of the histogram. Experimental results show that proposed scheme can hide a large amount of information within a video frame with a high degree of robustness against H.264/AVC encoding.

Study of histopathological cancerous tissue is one of the most reliable ways to grade various types of cancers. The result of grading helps the physicians to diagnose and prescribe suitable prognosis. The focus of this paper is on a CAD for automatic analysis of breast cancer histopathological Images to count mitosis as an important criteria for the breast cancer grading. To achieve this aim, sets of specific digital histopathological data are used which are captured by particular microscopic scanners named as Aperio XT and Hamamatsu NanoZoomer scanners. In the proposed method, these acquired images are employed and processed based on digital image processing approaches like 2-D anisotropic diffusion as a pre-process and morphological process. For extraction of pixel-wise features from predetermined mitotic regions, an statistical approach based on color information such as maximum likelihood estimation is employed. To prevent misclassification of mitosis and non-mitosis objects, an object-wise completed local binary pattern (CLBP) is proposed to extract texture features robust against rotation and color-level changes, and finally support vector machine (SVM) is used to classify the extracted feature vectors. Having computed the evaluation criteria, our proposed method performs better f-measure (70.94% for Aperio XT scanner images and 70.11% for Hamamatsu images) among the methods proposed by other participants at ICPR2012 Mitosis detection in breast cancer histopathological images.

Estimation of age and gender is one of the foremost challenges in computer vision and has lots of applications in recommender systems establishment, security and systems which deal with computer vision. In this paper, a method for age and gender estimation using facial images are proposed by focusing on the extraction of robust features existing in facial photos. The main stage for age estimation is done in two main steps. At the first step classification and extraction of the global features is done, and in the second step ratios which help distinguishing child (1 to 12 year-old children) from youth (13 to 40 year-old men) are described and in the next step by using the same procedure, seniors (from 41 to 80 years old) are separated from the two former groups. For gender estimation purpose, ratios computed in the previous step, are employed and finally the correct gender estimation is done.

A semi-fragile lossless data hiding (LDH) Scheme based on histogram distribution shift in integer wavelet transform (IWT) domain is proposed in this paper. In the proposed scheme, the transform approximation image is divided into non-overlapping blocks. In each block, the differences between the neighboring elements are computed and a histogram is made on the difference values. The secret data are embedded into the blocks based on a multi-level shifting mechanism of the histogram. The proposed method enables the exact recovery of the original host signal upon extracting the embedded information, if the watermarked image is not affected by any other process. The performance of proposed scheme is evaluated with respect to imperceptibility, robustness, and data payload capacity by simulations. Comparing with the state-of-the-art known techniques, the proposed method performs a higher performance.

Traffic control systems such as traffic lights play an inevitable role in the current world's transportation guidance and driving fluency. In this paper, an automatic traffic control system based on advance software architecture is proposed. In the proposed architecture, automatic car plate recognition and driver verification based on fingerprint biometric are mixed with each other. The license plate recognition part is adapted for Persian or Farsi characters. The persian or farsi characters recognition is done by a very simple Normalized cross correlation which is very analogous to Euclidian distance criterion. To improve the functionality of this platform, some special and innovative digital image processing are employed so that the program is able to extract and recognize the car plate and its related characters even if in the low light or shiny conditions. The fingerprint recognition system is also added to the proposed traffic control system to ensure the authority of the car driver to enter to the places with high privacy and security limitations. The simulation results demonstrate the efficiency and suitable performance of the proposed automatic traffic control system.

In this paper, a robust lossless data hiding scheme on uncompressed video data based on histogram distribution constrained (HDC) scheme is proposed. The proposed method is a reversible data hiding algorithm which enables exact recovery of the original video without any distortion after extracting hidden information, if the watermarked video is not affected by any other process. The algorithm is robust against H.246/AVC video compression. In the proposed method, the luminance components of video frames are used to embed the watermark data bits. The luminance component is divided into non-overlapping blocks and the arithmetic difference of each block is computed. In the next step, the hiding data are embedded into the blocks by shifting the arithmetic difference values. The performance of the proposed algorithm is evaluated with respect to imperceptibility, robustness against H.264 video coding and data payload capacity by simulations. Simulation results demonstrate a high performance for the proposed method.

The privacy and integrity protection of digital documents specially the case of biomedical images is a vital subject in the current world of telecommunication and digital multimedia exchanges. In this paper, a modified semi robust digital image watermarking method with tamper detection and recovery ability is proposed. For this purpose, the proposed method comprises two stages. In the first stage, a specific order of the integer wavelet coefficient of the original image forms the watermark which prepares both tamper detection and recovery ability. These abilities are created by a specific parsing of watermark bits named as dual watermarking. In the second stage of the proposed method, watermark embedding and extraction procedures are done. To be more robust, a Convolutional Error Correction Code is employed in the watermark establishment process. In the extraction process, the probable tamper will be detected and the original image shall be recovered. This method is also improved to be adopted for reversible ROI recovery of medical images even in the presence of some kind of intentional and unintentional attacks. The experimental results show that the proposed method has a high performance in the case of tamper detection and is able to recover original images from noisy ones.

There are many proposed methods for active sonar target echo detection in reverberation background via either optimal or suboptimal signal processing procedures. Among such methods, those ones which are based on mere filtering, e.g. matched filter, are both efficient and flexible, but suffer from some deficiencies such as high computational complexity or more additional requirements for post-processing. In this paper, an adaptively order-selected pre-whiten filtering method based on autoregressive (AR) modeling of the reverberation data at the active sonar receiving hydrophone is proposed. This method is able to overcome the deficiencies of former filtering methods. This is achieved by applying an AR pre-whiten filter that has its order selected adaptively using data partitioning. The adaptive order selection of AR pre-whiten filter is done by the use of FPEF which is a high performance AR order selection criterion. The results of simulation show that the proposed method is more efficient than the previously proposed order/reverse partition AR pre-whiten algorithm in the sense of echo-to-reverberation ratio (ERR).

Matching is an important step in any fingerprint recognition system. In this paper, a fingerprint matching technique based on hidden markov model is proposed. This method uses only the ridge orientation information around the reference point of registered fingerprint image. At the first step, fingerprint images are aligned according to a suitable reference point. Then in the second step, the ridge orientation field around this point is applied to the HMM based on an innovative topology. This topology provides many significant advantages such as simplicity, flexibility and generality. The selected orientation field forms specific feature vectors so that can be used for proposed HMM matching process. In the HMM matching process, the maximum likelihood between training and test feature vectors are tested according to a predefined threshold. For evaluating the proposed matching method, an artificial continuous classification had been applied over FVC2000 DB2_A. The results of experiment have proved the higher efficiency and robustness of the proposed method in comparison with the competing one.

As fingerprint verification depends strongly on the quality of fingerprint ridge orientation estimation, so the more accurate fingerprint ridge orientations estimate, the better verification will result in. In this paper, we have proposed a new technique for improving the coarse fingerprint orientation estimation smoothing using fingerprint orientation model based on 2D Fourier expansion with a special Modification on it (M-FOMFE). The modification we have used in this paper is taking into account the information of Coherence Matrix for fingerprint ridge orientation estimation. This matrix is used to retrieve the uncertainty of each ridge orientation block and improves the accuracy of old FOMFE coarse ridge orientation smoothing method. For evaluating the proposed method we use a fingerprint continuous classification system and also an exclusive one based on Poincare-Index algorithm for singular point detection. Compared to competing method, the experimental results show that the proposed method has better orientation estimation and classification results.