M. Schmidberger, U. Mansmann · 2008
Abstract. Studies of gene expression using high-density oligonucleotide microar-rays have become standard in a variety of biological contexts. The data recorded using the microarray technique are characterized by high levels of noise and bias. These failures have to be removed, therefore preprocessing of raw-data has been a research topic of high priority over the past few years. Actual research and computations are limited by the available computer hard-ware. For many researchers the available main memory limits the number of arrays that may be processed. Furthermore most of the existing preprocessing methods are very time consuming and therefore not useful for first and fast checks in laborato-ries. To solve these problems, the potential of parallel computing should be used. In microarray technologies and statistical computing parallel computing does not ap-pear to have been used extensively. For parallelization on multicomputers, message passing (MPI) methods and the R language will be used. Ideas for parallelization of VSN and FARMS as well as a large project in applied bioinformatics (> 5000 microarrays) will be discussed. Furthermore this presentation
Cited by 0
G. Tímár, Kristóf Karacs, Csaba Rekeczky · Journal of Circuits Systems and Computers · 2003
This report describes analogic algorithms used in the preprocessing and segmentation phase of offline handwriting recognition tasks. A segmentation-based handwriting recognition approach is discussed, i.e., the system attempts to segment the words into their constituent letters. In order to improve their speed, the utilized CNN algorithms, whenever possible, use dynamic, wave front propagation-based methods instead of relying on morphologic operators were embedded into iterative algorithms. The system first locates the handwritten lines in the page image, then corrects their skew as necessary. It then searches for the words within the lines and corrects the skew at the word level as well. A novel trigger wave-based word segmentation algorithm is presented, which operates on the skeletons of words. Sample results of experiments conducted on a database of 25 handwritten pages along with suggestions for future development are presented.
Cited by 10
Синицына, В. В. · 2025
Довольно небольшое количество существующих на сегодняшний день методов преобразуют цветовую информацию изображений в соответствии с корректным ее восприятием людьми с ахроматопсией. Однако проблема зрения ахроматопов заключается не только в невозможности отличать цвета, которые нормальный трихромат воспринимает различно, но и в преобразовании в данных методах всей цветовой составляющей изображения, в то время как часто возникает необходимость преобразовывать лишь некоторые области изображения, которые являются наиболее важными для ахроматопов с точки зрения ценности визуальной информации. Для решения данных задач предложен метод препроцессинга изображений с целью их последующей реколоризации для людей с ахроматопсией.
Cited by 0Open access
Baldemar Tene, D.C. Puchianu, N Angelescu · The Scientific Bulletin of Electrical Engineering Faculty · 2016
Abstract Digital mammograms are a useful tool for breast cancer detection. The quality of digital mammogram image can have a negative effect on the computer-aided diagnosis system. We investigate the use of automatic image thresholding methods for microcalcification detection in mammographic images. Experimental results on the BI-RADS 4 images dataset confirm the proposed approach.
Cited by 0Open access
Antonio Gomez-Skarmeta, Fernando Jiménez, Jesús Ibáñez et al. · 2007
Databases in the real world are often dynamic, incomplete, noisy, imprecise and very large and most of the problems which appear in the data mining step within the knowledge discovery process are due to an inadequate preprocessing of the data present in these databases. The use of techniques that can cope with the imprecision present in the data, represent a new approach which has been successful in dealing with problems of imprecise and uncertain knowledge in the context of complex systems. In this paper an hybrid learning systems, that combines fuzzy-evolutionary techniques in order to find useful features or variables to represent the data as a preprocessing step, is being investigated. Keywords: Data preprocessing, knowledge discovery, data mining, evolutionary algorithms, fuzzy sets. 1 Introduction Given a set of data for which we presume some functional dependency, the question arises whether there is a suitable methodology to derive (fuzzy) rules from these data that character...
Cited by 0
Diego Alberto Guevara Amaya, Jairo E. Serrano, Juan Carlos Martinez Santos et al. · 2025
Fog significantly degrades image quality by reducing contrast, detail, and color fidelity. This degradation negatively impacts computer vision tasks. This work presents a comparative evaluation of four preprocessing techniques for foggy image enhancement: Dark Channel Prior (DCP), CLAHE, Multi-Scale Retinex (MSR), and AOD-Net. The methods were tested on 61 images from the DAWN dataset. Three no-reference visual quality metrics (NIQE, BRISQUE, PIQE) and average processing time were used as evaluation criteria. Results show that CLAHE offers the best trade-off between quality and efficiency, making it suitable for real-time applications. AOD-Net demonstrates greater robustness and stability, but at a higher computational cost. In contrast, DCP and MSR exhibit limitations in either stability or execution time. The findings highlight that no single technique is universally optimal. The selection should depend on specific system requirements, balancing visual quality, robustness, and efficiency.
Cited by 0
Фролов, И. И., Перцев, Д. Ю., Куприянова, Д. В. · Репозиторий БГУИР (BSUIR Repository) · 2020
Проведен сравнительный анализ алгоритмов предварительной обработки, применяемые для сжатия гиперспектральных данных. Проанализированы достоинства и недостатки представленных подходов.
Cited by 0
Zhenghui Fang, Aifeng Liu, Ke Li et al. · 2024
In human pose estimation algorithms, the introduction of point cloud data allows for more accurate geometric information of the human body surface without the need for texture mapping. Point clouds are sparse, which effectively reduces storage and processing costs. However, point clouds also exhibit characteristics such as noise, data incompleteness, and processing complexity. To address these issues, a lightweight point cloud data preprocessing method is proposed in this paper. In this work, the background point cloud, ground point cloud, and human body point cloud are first segmented to obtain a clean human body point cloud dataset. Next, denoising techniques are applied to this dataset, removing random noise around the point sets. Finally, the point cloud is downsampled based on the number of points required for human skinning, thereby reducing the sample size and improving processing speed. Experimental results have verified that this method effectively improves the real-time performance and accuracy of human pose estimation algorithms using point clouds.
Cited by 0
Shapcott, Caroline Mary · Warwick Research Archive Portal (University of Warwick) · 2020
Pre-processing techniques were developed for cell identification algorithms. These algorithms which locate and classify cells in digital microscopy images are important in digital pathology. The pre-processing methods included image sampling and colour normalisation for standard Haemotoxilyn and Eosin (H&E) images and co-localisation algorithms for multiplexed images. Data studied in the thesis came from patients with colorectal cancer. Patient histology images came from `The Cancer Genome Atlas' (TCGA), a repository with contributions from many different institutional sites. The multiplexed images were created by TIS, the Toponome Imaging System. \n \nExperiments with image sampling were applied to TCGA diagnostic images. The effect of sample size and sampling policy were evaluated. TCGA images were also used in experiments with colour normalisation algorithms. For TIS multiplexed images, probabilistic graphical models were developed as well as clustering applications. NW-BHC, an extension to Bayesian Hierarchical Clustering, was developed and, for TIS antibodies, applied to TCGA expression data. Using image sampling with a sample size of 100 tiles gave accurate prediction results while being seven to nine times faster than processing the entire image. \n \nThe two most accurate colour normalisation methods were that of Macenko and a `Nave' algorithm. Accuracy varied by TCGA site, indicating that researchers should use several independent data sets when evaluating colour normalisation algorithms. Probabilistic graphical models, applied to multiplexed images, calculated links between pairs of antibodies. The application of clustering to cell nuclei resulted in two main groups, one associated with epithelial cells and the second associated with the stromal environment. For TCGA expression data and for several clustering metrics, NW-BHC improved on the standard EM algorithm. \n
Cited by 0
Joerg C. Lemm, Valeriu Beiu, John G. Taylor · 1995
Cited by 3
Israel Gohberg, Vadim Olshevsky · Journal of Complexity · 1994
Cited by 95
Kennedy Okokpujie, Etinosa Noma‐Osaghae, Samuel John et al. · Lecture notes in electrical engineering · 2017
Cited by 13
Söderhjelm, Staffan · Zenodo (CERN European Organization for Nuclear Research) · 1982
Technical note written as input to the early studies of the ESA space astrometry project Hipparcos.
Cited by 0Open access
Zhang Shi · 2003
The processing technology of Chinese characters images plays an important role in its recognition process.Here studies several algorithms for Chinese Characters image's preprocessing,especially pays more attention to the size binarization and gives all experimental results. The results indicate that the proposed algorithms not only are efficient and available but also build stable basis for the subsequent processing.
Cited by 0
Söderhjelm, Staffan · Zenodo (CERN European Organization for Nuclear Research) · 1982
Technical note written as input to the early studies of the ESA space astrometry project Hipparcos.
Cited by 0Open access
Markus Even · 2021
The past two decades have seen continuing progress in the field of deformation analysis based on Interferometric Synthetic Aperture Radar (InSAR) data. In particular, refined methods for processing distributed scatterers (DS) have been developed. A general idea is that in a pre-processing step DS can be transformed to quasi PS and that the quasi PS together with PS can be analyzed with an arbitrary PS algorithm. This can be achieved by grouping statistically similar pixels, estimating the coherence matrix based on the group of pixels and extracting the signal of the DS from the coherence matrix. In this study different alternatives for these three steps are investigated with help of comprehensive simulations. In particular, a powerful regularization method is introduced that has not been used in the InSAR context before. It improves considerable the results for those estimators of DS signal that make use of the inverse of the coherence matrix.
Cited by 4
R. Suguna, D. Sharmila · 2014
Abstract- Web logs take an important role to know about user behavior. Several pattern mining techniques were developed to understand the user behavior. A specific kind of preprocessing technique improves the quality and accuracy of the pattern mining algorithms. The existing algorithms have done the preprocessing activities for reducing the size of the log file and to identify the number of unique users and sessions. In order to identify the user interest level and group similar kind of users, User Interest Level Preprocessing (UILP) algorithm is newly proposed. This paper discusses the basics of web log preprocessing, existing preprocessing techniques, the proposed UILP algorithm and performance of the proposed UILP algorithm with existing algorithms to identify user interest level.
Cited by 8
Raphael Souza de Oliveira (12647110), Amilton Sales Reis Jr. (12647113), Erick Giovani Sperandio Nascimento (6146123) · Figshare · 2022
Cited by 0Open access
Orosz György · Pázmány Péter Katolikus Egyetem · 2015
Cited by 0Open access
György Orosz · REAL-PhD (Hungarian Academy of Sciences) · 2015
This thesis deals with text processing applications examining methods suitable for less-resourced and agglutinative languages, thus presenting accurate preprocessing algorithms.The first part of this study describes morphological tagging algorithms which can compute both the morpho-syntactic tags and lemmata of words accurately.A tool (called PurePos) was developed that was shown to produce precise annotations for Hungarian texts and also to serve as a good base for rule-based domain adaptation scenarios.Besides, we present a methodology for combining tagger systems raising the overall accuracy of Hungarian annotation systems.Next, an application of the presented tagger is described that aims to produce morphological annotation for speech transcripts, and thus, the first morphological disambiguation tool for spoken Hungarian is introduced.Following this, a method is described which utilizes the adapted PurePos system for estimating morpho-syntactic complexity of Hungarian speech transcripts automatically.The third part of the study deals with the preprocessing of electronic health records.On the one hand, a hybrid algorithm is presented for segmenting clinical texts into words and sentences accurately.On the other hand, domain-specific enhancements of PurePos are described showing that the resulting tagger has satisfactory performance on noisy medical records.Finally, the main results of this study are summarized by presenting the author's theses.Further on, applications of the methods presented are listed which aims less-resourced languages.
Cited by 1Open access