IJDAR Special Issue: Deep Learning for Doc. Rec. and Analysis
Job offers : Post-doc and PhD positions, Rouen (France)
Call for summer schools
Call for papers
Call for Datasets
Call for Contributions
Please contribute relevant news to TC10 groups.
Please send any relevant event, notice or link to the newsletter editor : Alicia Fornes.
Message from the editor
Welcome to the April edition of the TC10 newsletter.
I am Motoi Iwata, the new TC-10 newsletter editor, in substitution of Alicia Fornes.
I am an Associate Professor at Osaka Prefecture University. In Japan, cherry blossoms announce coming of Spring. I hope this newsletter will be one of such announcements for you. I look forward to serving the TC-10 community.
Please, send to me any relevant news to be shared among the TC10 community.
In this issue, you will find detailed information concerning ICDAR 2017, job offers, and the next call for papers. We are glad to announce that DAS 2018 will be held in Vienna, Austria.
We also include the preliminary GREC 2017 call for participation. This time, the GREC workshop will take place in Kyoto, together with other ICDAR workshops.
Dear TC10 members, Following the IAPR meetings in Cancùn, Mexico, during ICPR 2016, the TC-10 board members have changed. We have striven to balance senior and junior members and we have integrated a renewed team with young and dynamic individuals covering a broad range of scientific themes and geographic locations:
- Chair Dr. Bart Lamiroy, Université de Lorraine, Loria (UMR 7503), Nancy, France
- Vice Chair Dr. Alicia Fornés, Computer Vision Center, Universitat Autònoma de Barcelona, Spain
- Secretary and TC-10 Webmaster Prof. Jean-Christophe Burie, L3i, University of la Rochelle, France
- Educational officer Dr. Syed Saqib Bukhari, German Research Center for Artificial Intelligence - DFKI, Germany
- Newsletter Editor Dr. Motoi Iwata, Osaka Prefecture University , Sakai, Japan
- Dataset Curator Dr. Santosh K.C., University of South Dakota, USA
GREC workshops provide an excellent opportunity for researchers and practitioners at all levels of experience to meet colleagues and to share new ideas and knowledge about graphics recognition methods. Graphics Recognition is a subfield of document image analysis that deals with graphical entities in engineering drawings, comics, maps, architectural plans, musical scores, mathematical notation, tables, diagrams, etc.
The aim of this workshop is to maintain a very high level of interaction and creative discussions between participants, maintaining a "workshop" spirit, and not being tempted by a "mini-conference" model.
GREC 2017 will continue the tradition of past workshops held at the Penn State University (USA), Nancy (France), Jaipur (India), Kingston (Canada), Barcelona (Spain),, Hong Kong (China), Curitiba (Brazil), La Rochelle (France), Seoul (Corea), Lehigh (USA) and Nancy (France).
The workshop will comprise several sessions dedicated to specific topics related to graphics in document analysis and graphic recognition. For each session, there will be an invited presentation describing the state of the art and stating the open questions for the session's topic, followed by a number of short presentations that will contribute by proposing solutions to some of the questions or presenting results of the speaker's work. Each session will be concluded by a panel discussion.
Topics - Analysis and interpretation of graphical documents, such as: Engineering drawings, floorplans, mathematical expressions, comics, maps, music scores, patents, diagrams, charts, tables, etc. - Recognition of graphic elements, such as symbols, logos, stamps, dropcaps, drawings, etc. - Identification and localization of graphical mark-ups and annotations in written documents. - Raster-to-vector techniques. - Graphics-based information retrieval. - Historical graphics recognition and indexing. - Forensics (Writer identification/verification) in graphic documents. - Description of complete systems for interpretation of graphic documents. - Datasets and performance evaluation in graphics recognition. - Authoring, editing, storing and presentation systems for graphics multimedia documents. - 3-D models from multiple 2-D views (line drawings). - Digital ink processing. - Sketch recognition and understanding. - Camera-based graphics recognition. - Graphics recognition in born digital documents. - Analysis of graphics on new digital interfaces. - Graphics detection and recognition in real scenes.
The Second International Workshop on coMics ANalysis, Processing and Understanding (MANPU) November 10, 2017 Kyoto, Japan http://manpu2017.imlab.jp/
Comics is a medium constituted of images combined with text and other visual information in order to narrate a story. Nowadays, comic books are a widespread cultural expression all over the world. The market of comics continues to grow, for example, the market in Japan is about 4.25 billion USD in 2015. Moreover, from the research point of view, comics images are attractive targets because the structure of a comics page includes various elements (such as panels, speech balloons, captions, leading characters, and so on), the drawing of which depends on the style of the author and presents a large variability. Therefore comics image analysis is not a trivial problem and is still immature compared with other kinds of image analysis.
Topics - Comics Image Processing - Comics Analysis and Understanding - Comics Recognition - Comics Retrieval and Spotting - Comics Enrichment - Reading Behavior Analysis of Comics - Comics Generation - Copy protection - Fraud detection - Physical/Digital Comics Interfaces - Cognitive Processing and Comprehension of Comics - Linguistics Analysis of Comics
HDI 2017 will focus on how humans interact with written information around them, and the interfaces between users and documents. The term document is meant in the wider possible sense, to refer to any physical object that carries written information.
The workshop aims to create a space for debate between the Document Image Analysis and Recognition and the Human-Computer Interaction communities, and it is motivated by a number of recent developments that point to a coming change of paradigm in the way we interact with written information around us.
Fresh paper submissions as well as re-submissions of improved versions of rejected ICDAR papers (submitted along with the original reviews) are welcome. The topics of interest include among others: - Augmented documents - Linking physical and digital content - Reading behaviour analysis - Human factors - User experience and usability - Wearable sensors in reading - Interaction for facilitating Document Image Analysis (teaching reading systems through the way humans interact with written information) - Document Image Analysis for enabling Interaction - Content personalisation - Applications (e.g. document editing, interactive translation, collaborative editing)
Accepted papers will be published in the "ICDAR 2017 Workshops" IEEE CPS volume.
Important Dates Submission Deadline (normal papers): July 1, 2017 Submission Deadline (ICDAR resubmissions): July 15, 2017 Notification: July 31, 2017 Camera Ready: September 1, 2017 Workshop: November 10 or 11, 2017 (TBC)
Workshop Chairs Jean-Christophe Burie (L3i Laboratory, University of La Rochelle, France) MickaÃ«l Coustaty (L3i Laboratory, University of La Rochelle, France) Dimosthenis Karatzas (Computer Vision Centre, Spain) Koichi Kise (Osaka Prefecture University) Kai Kunze (Keio University, Japan)
ICDAR 2017: Call for Nominations: ICDAR 2017 Awards
NOMINATIONS DUE: May 15, 2017
The ICDAR Award Program is an established program designed to recognize individuals who have made outstanding contributions to the field of Document Analysis and Recognition in one or more of the following areas: - Research - Training of students - Research/Industry interaction - Service to the profession
Every two years, two awards categories are presented. Namely, the IAPR/ICDAR Young Investigator Award (less than 40 years old at the time the award is made), and the IAPR/ICDAR Outstanding Achievements Award. Each award will consist of a token gift and a suitably inscribed certificate. The recipient of the Outstanding Achievements award will be invited to give the opening key note speech at the ICDAR 2015 conference, introduced by the recipient from the previous conference.
Nominations are invited for the ICDAR 2017 Awards in both categories.
The nomination packet should include the following: 1. A nominating letter (1 page) including a brief citation to be included in the certificate. 2. A brief vitae (2 pages) of the nominee highlighting the accomplishments being recognized. 3. Supporting letters (1 page each) from 3 active researchers from at least 3 different countries.
A nomination is usually put forward by a researcher (preferably from a different Institution than the nominee) who is knowledgeable of the scientific achievements of the nominee, and who organizes letters of support.
Submission procedure is strictly confidential, and self-nominations are not allowed.
The deadline for receipt of nominations is May 15th, 2017 but early submissions are strongly encouraged.
The final decision will be made by the Awards Committee which is composed of the following members (ICDAR advisory board and previous awardees): Dimosthenis Karatzas, Bart Lamiroy, Dan Lopresti, Jean-marc Ogier, Josep Llados, Simone Marinai, Cheng-Lin Liu, Marcus Liwicki (2015 IAPR/ICDAR Young Investigator Award recipient), Venu Govindaraju (2015 IAPR/ICDAR Outstanding Achievements Awards recipient).
Nine workshops and three tutorials will be offered at ICDAR 2017. You can find the names and organizers for each below. Please direct any questions to the workshop and/or tutorial chairs (contact information is provided below).
Workshops - HDI 2017: First International Workshop on Human-Document Interaction Jean-Christophe Burie, Mickaël Coustaty, Dimosthenis Karatzas and Koichi Kise - CBDAR 2017: 7th International Workshop on Camera-Based Document Analysis and Recognition Lluis Gomez-Bigorda, Muhammad Muzzamil Luqman and Dimosthenis Karatzas - MANPU 2017: 2nd International Workshop on coMics Analysis, Processing and Understanding Jean-Christophe Burie, Toshihiko Yamasaki and Motoi Iwata - WML ICDAR: Workshop on Machine Learning Umapada Pal and Eric Granger - HIP 2017: 4th International Workshop on Historical Document Imaging and Processing Andreas Fischer, Angelika Garz, Kengo Terasawa and Bill Barrett - IWCDF 2017: 1st International Workshop on Computational Document Forensics Jean-Marc Ogier, Utpal Garain and Apostolos Antonacopoulos - ICDAR-OST: 1st International Workshop on Open Services and Tools for Document Analysis Marcel Würsch and Joseph Chazalon - GREC: 12th International Workshop on Graphics Recognition Alicia Fornés and Bart Lamiroy - MOCR: 6th International Workshop on Multilingual OCR Venu Govindaraju, Prem Natarajan, Santanu Chaudhury and Srirangaraj Setlur
Tutorials Each of tutorial will last half a day.
- RTRS: Building a Robust Text Reading System for Unconstrained Scene Images Zhizhong Su and Chengquan Zhang - GMPRDIA: Graph-based Methods in Pattern Recognition and Document Image Analysis Anjan Dutta, Muhammad Muzzamil Luqman, Josep Lladós Canet, Jean-Marc OGIER - WS: Word Spotting: From Bag-of-Features to Deep Learning Gernot A. Fink
We are pleased to announce that the ICDAR'17 that will be held in Kyoto(November, 10th-15th) will organize a set of competitions dedicated to a large set of document analysis problems. You are cordially invited to participate in this scientific event that will be a very good opportunity to objectively compare the quality of algorithms in different categories of challenges.
PLEASE CONSIDER SUBMITTING YOUR ALGORITHM(S) TO ALL COMPETITIONS IN THE SAME CATEGORY, as we asked the individual competition organizers to synchronize their evaluation metrics, if possible. You will find below the different categories of competitions, and the URL of their respective website, that will allow you to get all the required information for participating.
In 2011, the first Doctoral Consortium in the Document Analysis community was organized in conjunction with the International Conference on Document Analysis and Recognition (ICDAR). This has led to successful successor events at ICDAR 2013 and ICDAR 2015. The tradition of having a Doctoral Consortium as a satelite event to the ICDAR main conference will further be continued at ICDAR 2017 in Kyoto, Japan.
The goal of the ICDAR 2017 Doctoral Consortium is to create an opportunity for Ph.D. students to test their research ideas, present their current progress and future plans, and receive constructive criticism and insights related to their future work and career perspectives. A mentor (a senior researcher who is active in the field) will be assigned to each student to provide individual feedback. In addition, students will have the opportunity to present an overview of their research plan during a special poster session.
Participation in the ICDAR 2017 Doctoral Consortium will be limited to 25 students. Prospective participants are encouraged to submit their application by July 15 (submission procedure to be announced). The Doctoral Consortium Chairs will then review all applications received. Preference will be given to students who are at a stage in their studies most likely to benefit (i.e., they have identified a research direction and published some initial results, but the thesis is not yet set in stone).
Participation to the Doctoral Consortium will be free for all accepted students, i.e., there will be no extra registration fees!
The ICDAR 2017 Doctoral Consortium will take place the day before the main conference, i.e., on Sunday, November 12.
IMPORTANT DATES - Submission deadline: July 15 - Doctoral Consortium: November 12
We are looking forward to your participation!
Véronique Eglin and Rafael Dueire Lins TC-10/11 Educational Officers
ICFHR 2018: Moved to Niagara Falls, USA (repost)
After much discussion by the ICFHR 2018 conference chairs and with the approval of TC-11, the site for ICFHR 2018 has been moved from Rochester, NY (USA) to Niagara Fall, NY (USA). There are multiple reasons for this change; they include higher value for participants (e.g., better food, meeting rooms and surrounding environs) and Niagara being a more attractive meeting place for existing and potential members of the TC-11 community. The total cost to run the conference, as well as for individual participants should remain the same.
Any consortium interested in making a proposal to host an ICDAR should first familiarise themselves with the "Guidelines for Organizing and Bidding to Host ICDAR" document which is available on the TC10 and TC11 websites (http://iapr-tc10.univ-lr.fr
A link to the most current version of the guidelines appears below. Small updates in the guidelines are expected during the next few weeks, so please check on the Web site of TC11 for the latest version.
The submission of a bid implies full agreement with the rules and procedures outlined in that document. The submitted proposal must define clearly the items specified in the guidelines (Section 5.2).
It has been the tradition that the location of ICDAR conferences follows a rotating schedule among different continents. Hence, proposals from America are encouraged. However, high quality bids from other locations, for example, from countries where we have had no ICDAR before, will also be considered. Proposals will be examined by the ICDAR Advisory Board.
IEEE Transactions on Human-Machine Systems: Special Issue on Drawing and Handwriting Processing for User-Centered Systems
The April 2017 issue of IEEE Transactions on Human-Machine Systems is concerned with interactive systems that incorporate handwriting and/or drawing. The Special Issue aims to bring together some of the most valuable research contributions and applicable solutions in the field of drawing and handwriting processing for human–machine interaction.
Twelve papers are included in the special issue concerning diverse topics such as signature verification, user-centered document segmentation, sketch retrieval, combined speech and handwriting entry, emotion and handwriting, and dysgraphia.
The issue is available online now: THMS April 2017. Please spread the word, and take a moment to peruse the Table of Contents.
IJDAR Special Issue: Deep Learning for Document Analysis and Recognition
In recent years, applying Deep Learning techniques to Document Analysis and Recognition (DAR) tasks has produced significant performance gains. These have included impressive advances in handwriting recognition, word-spotting in document images, document segmentation, and text detection in natural scenes.
This special issue aims to report new advances in DAR using deep learning methods. Articles presenting reviews, perspectives, new methods and applications in DAR are cordially invited.
Topics of interest include, but are not limited to uses of Deep Learning for: - Document image processing and segmentation - Layout analysis - Character and text recognition - Scene text detection and recognition - Writer identification and signature analysis - Document retrieval - Context modeling - Graphics and symbol recognition - Other DAR tasks
Paper submissions are due AUGUST 31, 2017.
For further information, please see the Call for Submissions.
PhD position "Metric Learning on Graphs" - University of Rouen Normandy, France
PhD position "Metric Learning on Graphs" LITIS EA 4108 -- University of Rouen Normandy
Graphs have a high description power. Many pattern recognition methods rely on the notion of distance between objects. The Graph Edit Distance (GED) has an exponential computation complexity. This distance rely on cost functions associated to elementary edit operations (substitution, insertion, deletion of vertices or edges). Until now, these cost functions that depend on vertice/edge labels were given or determined with human expertise.
This PhD subject aims to transpose the ability of artificial deep neural networks to automatically determine relevant features and their weights, in the graph domain. The objective is to propose an approach that allows to determine cost functions of elementary edit operations used in the GED or its approximations through automatic learning. This automatic learning should be performed for several objective functions : - to respect at its best graph distance values given in a training dataset - to respect at its best node-to-node or subgraph-to-subgraph matchings.
These works will be conducted in the context of a collaboration with other computer science labs (LI Tours, GREYC Caen). Applications in chemo-informatics are considered with interactions with chemical science labs (LCMT Caen, CERMN Caen, COBRA Rouen)
Supervising team - Sébastien Adam (Professor at University de Rouen Normandy) - Pierre Héroux (Assistant Professor at University de Rouen Normandy) - Benoît Gaüzère (Assistant Professor INSA Rouen Normandy)
The applicant should have a theoretical knowledge in machine learning and graph theory. High programming skills are required. The selection will be done on the basis of the application form including a resume, marks, MSc marks, cover letter, recommendations, and then an interview.
They will be processed as soon as they are received, and a decision may be taken if a candidate meets the requirements.
Postdoctoral Position, Rouen, Normandy, France
Postdoctoral Position in Machine learning (LITIS, Université de Rouen, France)
The Machine Learning team at LITIS (France) is opening a postdoctoral position for a candidate with a strong experience in Machine Learning, with application to one of the following topics: Speech Processing, Handwriting Processing, Computer Vision, or Document Image Processing.
The candidate will be involved in pursuing the development of Handwriting recognition and document image processing technologies to extend the PlaIR platform that was originally developed for digitizing old Newspapers (http://www.plair.univ-rouen.fr). New contributions of the candidate are expected regarding Handwriting Recognition (HWR), by combining deep optical models with deep language models. He will contribute with other members of the team to various research projects in this domain.
SCIENTIFIC AND TECHNICAL SKILLS Experimented engineer or postdoctoral candidates having a strong background in one of the following topics : Neural Networks, Deep Neural Networks, Hidden Markov Models, Recurrent Neural Networks, statistical language models, computer vision.
The successful candidate should have advanced programming skills (Python, C/C++, Matlab). Experience of developing models within one of the popular frameworks such as Torch 7, TensorFLow, will be particularly appreciated.
APPLICATION Application should include a curriculum vita, a brief statement of research interests, and the names of at least three references.
We would like to remind you that the TC10 and TC11 Web sites always welcome contributions of new datasets or other resources related to the community. We would like to encourage all the TC10 and TC11 members to submit such material to the TC10 and TC11 for archiving. The availability of datasets, ground truth and performance evaluation tools online is not only good practice, but also a requirement for a field to progress.
We would like make a special request to the organizers of recent and future competitions. Independently of whether you have the competition datasets and evaluation tools available through other Web sites, please consider archiving them with TC10 / TC11 as well. Web sites often go off-line and useful resources are frequently lost forever.