Dr. Szilárd Vajda

Associate Professor

Placeholder Image

News

  • Recently, I was featured in an online publication in Romania, at Foter.ro. I am discussing research, education, ChatGPT and other things (in Hungarian). [04/16/2023]
  • I published a new Docker  Ubuntu Linux 22.04 LTS image for my CS students so that they can practice their Linux skills, programming skills (C, C++, Java, Python, Ruby, Lisp), debugging (gdb, valgrind), build system (make, cmake), full manual pages in English, etc. Just type docker search szilardvajda  or docker pull szilardvajda/ubuntu_2204lts_cs470 in your teminal to get it! If interested I can provide you with the corresponding Dockerfile as well! Feel free to use it! [Posted: 01/03/2023]
  • My students Lukas Keizur, Richard DeYoung, Kirsten Boyles, Graig Turnbull and Joe Corona were featured in Central's News. They built an app for electric vehicles [Demo] [Posted: 03/04/2022]
  • If you are insterested in doing Directed Research (see CS495) working with me, please contact me via e-mail. Possible topics: machine learning, medical/document/general image recognition, Alderaban robot programming, artificiall intelligence, deep neural networks, mobile app development, Alexa skills, etc. I am open to all kind of other topics as well. [Posted: 01/03/2022]
  • My graduate student, Hermann J. Nkouanga got the Graduate Student of the Year Award for his work in  "Automatic tuberculosis detection using chest x-ray analysis with enhanced structural information." [Presentation] [Graduate Scholar of the Year Award, SOURCE2020] [Posted: 08/13/2020]
  • Central Washington University ranked #339 on Money Magazine's Best Colleges for Your Money2019. [Posted: 08/12/2019]

Bio

Szilárd is currently a tenured faculty with the Computer Science Department at Central Washington University. He is conducting research in machine learning, pattern recognition, high-performance computing, etc. He is responsible for several graduate and undergraduate CS courses offered by the department. 

Previously he was a postdoctoral fellow with Lister Hill National Center for Biomedical Communication, National Library of Medicine, National Institutes of Health, Bethesda, MD. USA, where he was working on different image processing and recognition tasks involving medical documents, medical images and human faces. He also completed a postdoc at TU Dortmund, Dortmund, Germany. In the Pattern Recognition in Embedded System group he was involved in several document analysis projects. He also supervised graduate and undegraduate students. Previously, in Budapest, Hungary he was hired by Furukawa Electric Institute of Technology where he conducted research and development work for the automotive industry.

Szilárd completed his doctoral studies at READ group, Loria Research Center, University of Lorraine (former Henri Poincare University) in Nancy, France. He was interested in document analysis, especially Indian postal document recognition. 

Research

  • Szilárd is interested in a large variety of research topics evolving around machine learning, image recognition and document analysis. His favourite research topics are:
  • Document Analysis (postal documents, digital maps, bank cheques, forms, old manuscripts)
  • Handwriting Recognition (word /digit recognition in Roman/non-Roman based scripts, sign recognition using offline approaches) 
  • Neural Networks (deep networks, convolusional networks, training algorithms, outlier problem, network topology, parameter estimation/optimization)
  • Classifiers Combination (dynamic/static combination and fusion)
  • Machine Learning (convergence, boosting, pattern selection, variable selection, active learning, imbalanced dataset paradigm, supervised/non-supervised learning, Bayesian learning) 
  • Face Recognition (skin color detection, skin color classification, face detection, face matching, etc)
  • Medical Imaging (chest x-ray analysis, medical image modality detection)
  • Data Analytics (data visualization, automatic labeling, dimensionality reduction) 

Publications

 

Peer-reviewed Journal Papers

  • Szilárd Vajda, Alexandros Karargyris, Stefan Jäger, K. C. Santosh, Sema Candemir, Zhiyun Xue, Sameer K. Antani, George R. Thoma: Feature Selection for Automatic Tuberculosis Screening in Frontal Chest Radiographs. J. Medical Systems 42(8): 146:1-146:11 (2018). Impact factor: 2.098
  • Eugene Borovikov, Szilárd Vajda, Michael Bonifant, and Michael Gill. Looking at faces in the wild. Procedia Computer Science (Special Issue), 123:104 – 109, 2018. 8th Annual International Conference on Biologically Inspired Cognitive Architectures, BICA 2017 (Eighth Annual Meeting of the BICA Society), held August 1-6, 2017 in Moscow, Russia.
  • Eugene Borovikov, Szilárd Vajda, and Michael Gill. Face match for family reunification: Real-world face image retrieval. IJCVIP, 7(2):19–35, 2017. Citation factor: 7.484.
  • K. C. Santosh, Szilárd Vajda, Sameer K. Antani, and George R. Thoma. Edge map analysis in chest x-rays for automatic pulmonary abnormality screening. International Journal of Computer Assisted Radiology and Surgery, 11(9):1637–1646, 2016. Impact factor: 1.863.
  • Alexandros Karargyris, Jenifer Siegelman, Dimitris Tzortzis, Stefan Jaeger, Sema Candemir, Zhiyun Xue, K. C. Santosh, Szilárd Vajda, Sameer K. Antani, Les Folio, and George R. Thoma. Combination of texture and shape features to detect pulmonary abnormalities in digital chest x-rays. International Journal Computer Assisted Radiology and Surgery, 11(1):99–106, 2016. Impact factor: 1.863.
  • Szilárd Vajda, Daekeun You, Sameer Antani, and George Thoma. Large image modality labeling initiative using semi-supervised and optimized clustering. International Journal of Multimedia Information Retrieval, 4(2):143–151, 2015. Impact factor: 0.141.
  • Szilárd Vajda, Yves Rangoni, and Hubert Cecotti. Semi-automatic ground truth generation using unsupervised clustering and limited manual labeling: Application to handwritten character recognition. Pattern Recognition Letters, 58:23–28, 2015. Impact factor: 1.586.
  • Szilárd Vajda, Thomas Plötz, and Gernot Fink. Camera-based whiteboard reading for understanding mind maps. International Journal of Pattern Recognition and Artificial Intelligence, 29(3), 2015. Impact factor: 0.994.
  • Jan Richarz, Szilárd Vajda, Rene Grzeszick, and Gernot A. Fink. Semi-supervised learning for character recognition in historical archive documents. Pattern Recognition, 47(3):1011–1020, 2014. Impact factor: 3.399.
  • Yves Rangoni, Abdel Belaïd, and Szilárd Vajda. Labeling logical structures of document images using a dynamic perceptive neural network. International Journal of Document Analysis and Recognition, 15(1):45–55, 2012. Impact factor: 0.902.
  • Szilárd Vajda, Kaushik Roy, Umapada Pal, Bidyut B. Chaudhuri, and Abdel Belaïd. Automation of Indian postal documents written in Bangla and English, International Journal of Pattern Recognition and Artificial Intelligence, 23(8):1599–1632, 2009. Impact factor: 0.994.
  • Szilárd Vajda, Thomas Plötz, and Gernot A. Fink. Layout analysis for camera-based whiteboard notes. Journal of Universal Computer Science, 15(18):3307–3324, 2009. Impact factor: 0.466.

 

Book Chapters

  • Eugene Borovikov, Szilárd Vajda, Girish Lingappa, and Michael C. Bonifant. Parallel computing in face image retrieval: Practical approach to the real-world image search. In Mohan S. and Vani V., editors, Multi-Core Computer Vision and Image Processing for Intelligent Applications, pages 155–189. IGI Global, 2017.
  • Szilárd Vajda, Leonard Rothacker, and Gernot A. Fink. A method for camera-based interactive whiteboard reading system. In Camera-Based Document Analysis and Recognition, volume 7139 of Lecture Notes in Computer Science, pages 112–125. Springer, 2012.
  • Christian Kleine-Cosack, Marius H. Hennecke, Szilárd Vajda, and Gernot A. Fink. Exploiting acoustic source localization for context classification in smart environments. In Ambient Intelligence, volume 6439 of Lecture Notes in Computer Science, pages 157–166. Springer Berlin / Heidelberg, November 2010.
  • Hubert Cecotti, Szilárd Vajda, and Abdel Belaïd. High performance classifiers combination for handwritten digit recognition. In Sameer Singh, Maneesha Singh, Chidan and Apté, and Petra Perner, editors, Pattern Recognition and Data Mining, Third International Conference on Advances in Pattern Recognition, ICAPR 2005, Bath, UK, August 22-25, 2005, Proceedings, Part I, volume 3686 of Lecture Notes in Computer Science, pages 619–626. Springer, 2005.

 

Peer-reviewed conference papers

  • Hermann Jepdjio and Szilárd Vajda.  A fast and efficient k-nearest neighbor classifier using a convex envelop.  In International Conference on Recent Trends in Image Processing and Pattern Recognition, (RTIP2R), Communications in Computer and Information Science. Springer, 2021
  • Hermann Y. Nkouanga and Szilárd Vajda. Automatic Tuberculosis Detection Using Chest X-ray Analysis With Position Enhanced Structural Information. In IEEE International Conference on Pattern Recognition. IEEE Computer Society, 2020.
  • Dmytro Dovhalets, Boris Kovalerchuk, Szilárd Vajda and Razvan Andonie. Deep Learning of 2-D Images Representing n-D Data in General Line Coordinates, 4th International Symposium on Affective Science and Engineering, Japan Society of Kansei Engineering, Wanatchee, WA, USA, 2018.
  • Dmytro Dovhalets and Szilárd Vajda. A costophrenic angle estimator in frontal chest radiographs. In IEEE Conference on Biomedical and Health Informatics. IEEE Computational Intelligence Society, 2018.
  • Szilárd Vajda and Santosh KC. A fast k-nearest neighbor classifier using unsupervised clustering. In International Conference on Recent Trends in Image Processing & Pattern Recognition Bidar, India, December 16-17, 2016., Communications in Computer and Information Science. Springer, 2016.
  • Eugene Borovikov and Szilárd Vajda. Facematch: real-world face image retrieval. In International Conference on Recent Trends in Image Processing & Pattern Recognition Bidar, India, December 16-17, 2016., Communications in Computer and Information Science. Springer, 2016.
  • Razvan Andonie, Anne M. Johansen, Amy L. Mumma, Holly C. Pinkart, and Szilárd Vajda. Cost efficient prediction of cabernet sauvignon wine quality. In IEEE Symposium Series on Computational Intelligence. IEEE Computational Intelligence Society, 2016.
  • KC Santosh, Szilárd Vajda, Sameer Antani, and George Thoma. Automatic pulmonary abnormality screening using thoracic edge map. In 28th Int. Symposium on Computerbased Medical Systems, pages 360–361. IEEE Computer Society, 2015.
  • Szilárd Vajda, Daekeun You, Sameer Antani, and George Thoma. Label many with a few: Semi-automatic medical image modality discovery in a large image collection. In IEEE Symposium Series on Computational Intelligence, pages 167–173, Orlando, FL, USA, 2014. IEEE.
  • Szilárd Vajda and Barna Szocs. A neural network based distance function for the k-nearest neighbor classifier. In International Conference on Frontiers in Handwriting Recognition, pages 429–433, Crete, Greece, 2014. IEEE Computer Society.
  • Akmal Junaidi, René Grzeszick, Gernot A. Fink, and Szilárd Vajda. Statistical modeling of the relation between characters and diacritics in Lampung script. In International Conference on Document Analysis and Recognition, pages 663–667. IAPR, IEEE Computer Society, 2013.
  • Hubert Cecotti and Szilárd Vajda. A radial neural convolutional layer for multi-oriented character recognition. In International Conference on Document Analysis and Recognition, pages 668–672. IAPR, IEEE Computer Society, 2013.
  • Hubert Cecotti and Szilárd Vajda. Multi-class rejection evaluation in handwritten character recognition. In International Conference on Document Analysis and Recognition, pages 445–449. IAPR, IEEE Computer Society, 2013.
  • Eugene Borovikov, Szilárd Vajda, Girish Lingappa, Sameer Antani, and George Thoma. Face matching for post-disaster family reunification. In International Conference on Healthcare Informatics, pages 131–140. IEEE Computer Society, 2013.
  • Barna Szocs, Szilárd Vajda, and Judit Robu. D.A.C. draw and calc - the intuitive calculator. In 10th Jubilee International Symposium on Intelligent Systems and Informatics, pages 157–163, Subotica, Serbia, 2012. IEEE Computer Society.
  • Leonard Rothacker, Szilárd Vajda, and Gernot. A. Fink. Bag-of-feature representations for offline handwriting recognition applied to Arabic script. In International Conference on Frontiers in Handwriting Recognition, pages 149–154, Bari, Italy, 2012. IEEE Computer Society.
  • Jan Richarz, Szilárd Vajda, and Gernot A. Fink. Towards semi-supervised transcription of handwritten historical weather reports. In International Workshop on Document Analysis Systems, pages 180–184, Gold Coast, Queensland, Australia, 2012. IEEE Computer Society.
  • Jan Richarz, Szilárd Vajda, and Gernot. A. Fink. Annotating handwritten characters with minimal human involvement in a semi-supervised learning strategy. In International Conference on Frontiers in Handwriting Recognition, pages 23–28, Bari, Italy, 2012. IEEE Computer Society.
  • Szilárd Vajda, Akmal Junaidi, and Gernot A. Fink. A semi-supervised ensemble learning approach for character labeling with minimal human effort. In International Conference on Document Analysis and Recognition, pages 259–263. IAPR, IEEE Computer Society, 2011.
  • Szilárd Vajda and Gernot A. Fink. Strategies for training robust neural network based digit recognizers on unbalanced data sets. In International Conference on Frontiers in Handwriting Recognition, pages 148–153, Kolkata, India, November 2010. IAPR, IEEE Computer Society.
  • Szilárd Vajda and Gernot A. Fink. Exploring pattern selection strategies for fast neural network training. In International Conference on Pattern Recognition, pages 2913–2916, Istanbul, Turkey, 2010. IAPR, IEEE Computer Society.
  • Gernot A. Fink, Szilárd Vajda, Ujjwal Bhattacharya, Swapan K. Parui, and Bidyut. B. Chaudhuri. Online Bangla word recognition using sub-stroke level features and hidden Markov models. In International Conference on Frontiers in Handwriting Recognition, pages 393–398, Kolkata, India, 2010.
  • Szilárd Vajda and Abdel Belaïd. Structural information implant in a context based segmentation-free HMM handwritten word recognition system for Latin and Bangla script. In Proc. Int. Conf. on Document Analysis and Recognition, pages 1126–1130, Seoul, Korea, 2005. IEEE Computer Society.
  • Kaushik Roy, Szilárd Vajda, Umapada Pal, Bidyut B. Chaudhuri, and Abdel Belaïd. A system for Indian postal automation. In Proc. Int. Conf. on Document Analysis and Recognition, pages 1060–1064, Seoul, Korea, 2005. IEEE Computer Society.


Peer-reviewed workshop papers

  • Szilárd Vajda, Sameer K. Antani, and George R. Thoma. National library of medicine (NLM) at imageclef2015: Medical clustering task. In Working Notes of CLEF 2015 - Conference and Labs of the Evaluation forum, Toulouse, France, September 8-11, 2015., 2015.
  • KC Santosh, Szilárd Vajda, Sameer Antani, and George Thoma. Automatic pulmonary abnormality screening using thoracic edge map. In 28th Int. Symposium on Computer based Medical Systems, pages 360–361. IEEE Computer Society, 2015.
  • Yves Rangoni, Eric Ras, and Szilárd Vajda. Using handwriting recognition modality in tangible user interface. In IAPR International Workshop on Graphics Recognition (GREC). 2013.
  • Szilárd Vajda, Leonard Rothacker, and Gernot A. Fink. A camera-based interactive whiteboard reading system. In International Workshop on Camera-Based Document Analysis and Recognition, pages 91–96. IAPR, 2011.
  • Akmal Junaidi, Szilárd Vajda, and Gernot A. Fink. Lampung - a new handwritten character benchmark: Database, Labeling and Recognition. In International Workshop on Multilingual OCR, pages 105–112. IAPR, ACM, 2011. Best Paper Award.
  • Nils Y. Hammerla, Thomas Plötz, Szilárd Vajda, and Gernot A. Fink. Towards feature learning for HMM-based offline handwriting recognition. In International Workshop on Frontiers of Arabic Handwriting Recognition, Istanbul, Turkey, 2010. IAPR.
  • Szilárd Vajda, Tobias Ramforth, Thomas Plötz, and Gernot A. Fink. Camera-based analysis of whiteboard notes. In 3rd Int. Workshop on Camera-Based Document Analysis and Recognition, pages 42–49, Barcelona, Spain, 2009.
  • Szilárd Vajda, Hubert Cecotti, Yves Rangoni, and Abdel Belaïd. A fast learning strategy using pattern selection for feedforward neural networks. In International Workshop on Frontiers in Handwriting Recognition, La Baule, France, pages 145–150, 2006.
  • Szilárd Vajda and Abdel Belaïd. How to speed up the learning mechanism in a connectionist model. In IAPR International Workshop on Neural Networks and Learning in Document Analysis and Recognition, Seoul, Korea, pages 13–17, 2005.
  • Kaushik Roy, Szilárd Vajda, Umapada Pal, Bidyut B. Chaudhuri, and Abdel Belaïd. A system for Indian postal automation. In International Workshop on Document Analysis, Kolkata, India, 2005.
  • Szilárd Vajda, Hubert Cecotti and Abdel Belaïd. Neural and stochastic classifiers combination for handwritten digits characters. In IAPR International Workshop on Neural Networks and Learning in Document Analysis and Recognition, Seoul, Korea, 2005.
  • Hubert Cecotti, Szilárd Vajda, and Abdel Belaïd. HMM based viterbi paths for rejection correction in a convolutional neural networ classifier. In IAPR International Workshop on Neural Networks and Learning in Document Analysis and Recognition, 2005.
  • Kaushik Roy, Szilárd Vajda, Umapada Pal, and Bidyut B. Chaudhuri. A system towards Indian postal automation. In International Workshop on Frontiers in Handwriting Recognition, Tokyo, Japan, pages 580–585. IEEE Computer Society, 2004.
  • Ujjwal Bhattacharya, Szilárd Vajda, Anirban Mallick, Bidyut Baran Chaudhuri, and Abdel Belaïd. On the choice of training set, architecture and combination rule of multiple MLP classifiers for multiresolution recognition of handwritten characters. In International Workshop on Frontiers in Handwriting Recognition, Tokyo, Japan, pages 419–424. IEEE Computer Society, 2004.


Technical reports

  • Szilárd Vajda. Training and evaluation of the models for isolated character recognition. Technical report, Inria, Nancy, France, 2002.
  • Szilárd Vajda. Classifiers combination for recognition score improvement. Technical report, Inria, Nancy, France, 2002.
  • Szilárd Vajda. Characterization and normalization of the image database. Technical report, Inria, Nancy, France, 2002.


Others

Teaching

Current Students

Jia Song, MS in Computer Science (ongoing), Thesis: Chinese Handwritten Character Recognition, Computer Science Department, Central Washington University, Ellensburg, WA, USA, 2018

Former students

  • Su Chen, MS in Computer Science (ongoing), Thesis: Coocoo a WebAssembly Compiler for Image and 3D Material Generation, Computer Science Department, Central Washington University, Ellensburg, WA, USA, 2021
  • Hermann Jepdjio Nkouanga, MS in  Computer Science, Thesis: Chest x-ray segmentation and recognition using machine learning strategies, Computer Science Department, Central Washington University, Ellensburg, WA, USA, 2020
  • Brian Hooper, MS in Computer Science, Thesis: Medical image recognition using deep neural networks, Computer Science Department, Central Washington University, Ellensburg, WA, USA, 2020
  • Cole Webb, Undergraduate research, Topic: Whiteboard Reading with Rasberry Pi, Computer Science Department, Central Washington University, Ellensburg, WA, USA, 2018-2019
  • Dmytro Dovhalets, Graduate research, Topic: Costrophrenic Angle Estimation in Chest X-Rays, Computer Science Department, Central Washington University, Ellensburg, WA, USA, 2018 (research paper)
  • Yishui Liu, MS in Computer Science, Thesis: Traffic Sign Recognition, Computer Science Department, Central Washington University, Ellensburg, WA, USA, 2017 (Outstanding Oral Presentation Award, SOURCE)
  • Dipayan Banik, MS in Computer Science, Thesis: User Authentication using Doodle Recognition for Smartphone Application, Central Washington University, Ellensburg, WA, USA, 2017
  • Barna Szőcs, MS in Computer Science, Thesis: Neural Network Based Distance Functions, Computer Science Department, Babes Bolyai University, Cluj-Napoca, Romania, 2014 (research paper)
  • Barna Szőcs, B.Sc. in Computer Science, Thesis: The Intuitive Calculator, Computer Science Department, Babes Bolyai University, Cluj-Napoca, Romania, 2012 (research paper)
  • Leonard Rotchacker, M.Sc. in Computer Science, Thesis: Learning Bag-of-Features Representations for Handwriting Recognition, Computer Science Department, Technical University of Dortmund, Dortmund, Germany (jointly with Prof. Gernot A. Fink), 2011
  • Katrin Erlinghagen, B.Sc. in Computer Science, Thesis: Automatic Reading of Business Cards Using a Camera-Phone, Computer Science Department, Technical University of Dortmund, Dortmund, Germany (jointly with Prof. Gernot A. Fink), 2011
  • Akmal Junaidi, Ph.D in Computer Science, Computer Science Department, Technical University of Dortmund, Dortmund, Germany (jointly with Prof. Gernot A. Fink), 2010

 

Former interns

 

 

Topics

  • Medical image analysis (classification of healthy/unhealthy lungs, costophrenic angle estimation in chest x-rays)
  • Old documents segmentation (separating old handwriting in old medieval documents)
  • Alderaban robot programming (man-machine interaction, machine-machine interaction)
  • Drone programming (man-machine interaction)
  • K-Nearest neighbor classifier improvement (speed-up existing solutions)
  • Handwriting recognition (English, Arabic, Chinese, Bangla, Lampung, etc.)
  • Character and digit recognition
  • Automatic image coloring using machine learning
  • Business card recognition (mobile application development for iPhone or Android using OCR)
  • Licence plate recognition (mobile application development for iPhone or Android using OCR)
  • Bring your own idea (I am open for all kind of new ideas, technologies, solutions, frameworks, etc.)

AY2022-2023

  • SafeCampus, app to help Central to have a better response for emergency situations. [Demo]
  • Kare, home monitoring app for medically complex children. [Demo]
  • EV Charger, app to make charging easier and fun. [Demo] - in collaboration with Envorso.
  • LoadUp, logistic dispatch software. [Demo] -in collaboration with Taban Cosmos.
  • Rental Data Aggregation, app to handle rental data information. [Demo] - in collaboration with Taban Cosmos.
  • Receipt Reader, an app to handle receipts. [Demo] -in collaboration with Burhan Ul Hag.

AY2021-2022

  • Anti-Ebay website to do electronic business differently. [Demo]
  • Digital Pantry to handle your pantry items in your phone. [Demo]
  • EV Charging app to handle electric charging stations. [Demo] - in collaboration with Envorse.

AY2020-2021

  • Guide Recommender [Demo]
  • XuriBot [Demo]

AY2019-2020

AY2018-2019

Miscellaneous

Contact

Samuelson 216-E