IAPR TC11 Newsletter 2018 8

From TC11
Revision as of 07:43, 29 August 2018 by Zanibbi (talk | contribs)

(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)
Jump to: navigation, search
IAPR logo.gif Tc-11 Logo v2 72dpi.png

August, 2018



Click on the buttons below to view sections of the newsletter.

  • Message from the Editor
  • Report: ICFHR 2018 (Niagara Falls, USA)
  • Report: The 2nd IAPR TC10/TC11 Summer School (La Rochelle, France; repost)
  • Dates and Deadlines
    • Deadlines
    • Upcoming Conferences and Events
  • Calls for Papers
    • ICDAR-IJDAR Journal Track (New for ICDAR 2019)
      • Format
      • Procedure and Deadlines
      • Q&A
    • CFP: International Workshop on Robust Reading (IWRR - repost)
  • IJDAR
    • IJDAR: Latest Issue (Vol. 21, Issue 3)
      • Special Issue on Deep Learning for Document Analysis and Recognition
    • IJDAR Discount for IAPR Members (repost)
  • Datasets (repost)
  • Careers
    • IRISA/INSA Rennes (France): Research Engineer/PostDoc Position (3 Years - repost)
    • Student Industrial Internship Opportunities (IAPR - repost)



August was a busy month for the TC-11 community. In early August we had the 16th International Conference on Frontiers in Handwriting Recognition (ICFHR), and last week ICPR was held in Beijing. A report on ICFHR is included in this edition of the newsletter, and a report from ICPR will be included in a future edition.

Please note that the regular paper submission deadline for ICDAR 2019 has been moved forward to February 2019. For this ICDAR there is a new joint IJDAR-ICDAR track, where one submits a paper to IJDAR, and if accepted the paper appears in IJDAR and is also presented at ICDAR 2019. The submission deadline for the IJDAR-ICDAR track is much earlier (November). Please start planning now for how you will submit your work to our flagship conference, which will be held in Sydney, Australia.

Speaking of IJDAR, the new Special Issue on Deep Learning for Document Analysis and Recognition (Vol. 21, No. 3) is now available. The table of contents and links to papers are also included in the newsletter.

The International Workshop on Robust Reading (IWRR) also has an upcoming deadline in late September.

Best of luck to all of you who are preparing for or have already started a new academic year. Take care, and talk to you soon.

Join us! If you are not already a member of the TC-11 community, please consider joining the TC11 mailing list. Follow us on Twitter (iapr_tc11): https://twitter.com/iapr_tc11

Richard Zanibbi, TC-11 Communications Officer
( rxzvcs@rit.edu )

Icfhr2018.jpg

Icpr2018.png



The 16th International Conference on Frontiers in Handwriting Recognition was held in Niagara Falls, USA from August 5th through August 8th.

Over 130 people attended the conference, which combined a strong technical program, an attractive and comfortable meeting space with great food (the Niagara Falls Event and Convention Center), and the natural beauty of Niagara Falls. The banquet was held outdoors at the Old Fort Niagara, which was first used by the French military in the late 1600s.

ICFHR Banquet Photo.png

There was also plenty of fun to be had, particularly around the falls. Pictures from the conference are available through flickr.

Technical Program. This was the first ICFHR to utilize a double-blind reviewing process, and we introduced ‘Birds-of-a-Feather’ talks, which were informal discussions on specific topics held during lunch, led by 1-2 session leaders. To our pleasant surprise, these informal meetings were so successful that the most common complaint was a lack of space at the tables where the discussions were held.

There were 125 paper submissions from 29 countries, of which 32 were accepted for Oral presentation (26%), and 65 were accepted for Poster presentation (52%). The acceptance rate was 97/125 = 78%. The number of submissions was comparable to previous ICFHRs. All papers received at least 2 reviews, with 105 of them (84%) receiving three or more reviews.

Some very interesting keynote talks were given by Kevin Knight and Gregory R. Crane, addressing handwriting recognition from the different perspectives of a leading Computer Scientist and a leader in the Digital Humanities.

The program is available online, along with posters and slides provided by some of the authors. The proceedings will be available through IEEE Xplore in the coming weeks.

Sponsorship. The conference secured both sponsorship and participation from leading companies creating handwriting recognition-based products (MyScript, Wiris, and Hyperscience), USC/ISI, and leaders in the broader computing space (Google and Apple). An industrial panel was held by David Doermann on the last day of the conference, exploring handwriting recognition technologies in industry, and common interests between academic and industrial researchers working with handwriting. Also, thanks to support from the IAPR, we were able to provide funds for the IAPR ICFHR 2018 Awards.

Awards. The IAPR ICFHR 2018 Award recipients were:

  • Best Paper (500 USD): Kenny Davila and Richard Zanibbi, for “Visual Search Engine for Handwritten and Typeset Math in Lecture Videos and LaTeX Notes”
  • Best Student Paper (300 USD): Eugen Rusakov, Leonard Rothacker, Hyunho Mo, and Gernot A. Fink, for “A Probabilistic Retrieval Model for Word Spotting based on Direct Attribute Prediction”
  • Best Poster (200 USD): Harald Scheidl, Stefan Fiel, and Robert Sablatnig, for “Word Beam Search: A Connectionist Temporal Classification Decoding Algorithm”

Our sincerest thanks to all the participants, Organizing Committee members, support staff and student volunteers for making ICFHR 2018 a great success. Hope to see many of you at ICFHR 2020 in Dortmund!

Richard Zanibbi, General Co-Chair, ICFHR 2018
( rxzvcs@rit.edu )



The 2nd IAPR TC-10/TC-11 Summer School on Document Analysis (SSDA) was held in La Rochelle, from July 2nd to 6th 2018.

SSDA2018.jpg


The aim of this summer school was to give to new students in the field of DAR (Document Analysis and Recognition) an overview of all the traditional approaches to process and analyse documents but also to focus on new trends such as deep learning, the use of interactive devices, the fraud issues in documents, and so on. Different talks were given by international researchers from France, Switzerland, Spain, Japan and the United States.

In addition to traditional oral talks, three practical sessions were carried out in order to apply some techniques addressed during the talks. The interactive sessions (such as posters and self-introduction session) were a good environment for exchanges between participants and senior researchers. Finally, a visit of the L3i laboratory was given, including demonstrations of research activities in the field of DAR.

Two awards were given. The poster award was given to Florian Westphal (Blekinge Institute of Technology, Sweden) for his poster titled “Efficient Binarization using Heterogeneous Computing and Interactive Learning”. The excellence award was given to Tarin Clanuwat (Center for Open Data in the Humanities, Tokyo, Japan) for her outstanding participation.

This event gathered PhD students and junior researchers from 11 countries:

  • France: 7
  • Vietnam: 4
  • Sweden: 2
  • Finland: 2
  • Germany: 2
  • Indonesia: 2
  • Pakistan: 1
  • Austria: 1
  • Tunisia: 1
  • Italy: 1
  • Japan: 1

Thanks to the support of the IAPR, some travel grants and fee waivers were provided for participants with limited resources. Moreover, 17 master students (16 from China and 1 from Brazil) attended this summer school in order to discover the field of Document Analysis.

Jean-Christophe Burie, SSDA General Chair
( jcburie@univ-lr.fr )



Deadlines

2018

2019

Upcoming Conferences and Events

2018

2019 and Later

  • ICDAR 2019, Sydney, Australia (September 22-25, 2019)
  • ICFHR 2020. Dortmund, Germany (September 8-10, 2020)
  • ICFHR 2022. Hyderabad, India (December, 2022)



ICDAR-IJDAR Journal Track (New for ICDAR 2019)

New: IJDAR Journal Track at ICDAR 2019

For the first time, ICDAR 2019 (http://www.icdar2019.org) is opening a journal track that will offer the benefit of the rapid turnaround and dissemination times of a conference while providing the paper length, scientific rigor, and careful review process of an archival journal.

Important Dates

Nov. 15, 2018:  Initial submission deadline
Jan. 15, 2019:  Initial decisions (accept, minor/major revision, reject)
Feb. 15, 2019:  ICDAR regular paper submission deadline 
Apr. 15, 2019:  Submission deadline of the revised version 
                of IJDAR papers
 May 30, 2019:  Final decisions (accept to journal track, 
                or move to journal-only process)

The ICDAR-IJDAR journal track invites high-quality submissions that present original work in the areas of Document Analysis and Recognition appropriate to both the International Conference on Document Analysis and Recognition (ICDAR) and the International Journal on Document Analysis and Recognition (IJDAR, Impact Factor: 1.298). Accepted papers will be published in a special issue of IJDAR, and will receive an oral presentation slot at the ICDAR 2019 conference.

Journal versions of previously published conference papers or survey papers will not be considered for this special issue. Such submissions can be submitted as journal-only papers via the regular IJDAR procedures.

Authors who submit their work to the journal track commit themselves to present their results at the ICDAR conference in case of acceptance. Springer-Nature, the publisher of IJDAR, will make the papers accepted for the journal track freely available in a time frame of four weeks around the conference, beyond being available in the archival journal.

Format

Papers submitted to the journal track should follow the format of IJDAR submissions and all of the standard guidelines (https://www.springer.com/computer/image+processing/journal/10032).

Procedure and Deadlines

The deadline for journal track submissions is November 15th 2018. Earlier submissions are strongly encouraged in order to facilitate the reviewing process that will follow the IJDAR standards. A first decision about the papers should be released to the authors by January 15th, 2019.

Accepted papers will be scheduled for publication in the special issue and for presentation at ICDAR 2019. The authors of rejected papers are encouraged to submit their paper to ICDAR, following the conference guidelines.

Authors of papers accepted with minor or major revision should submit the revised version of their paper by April 15th 2019. However, if the revised paper does not fulfill the recommendations of reviewers, it will be forwarded to the regular IJDAR reviewing pipeline and will not be presented at ICDAR.

The final decision about the journal track papers should be released before May 30th 2019. Papers still under review at this date will remain in the reviewing pipeline of the journal. The deadline for camera ready versions of the accepted papers in the journal track will be June 25th and the expected publication date of the Special Issue is September 5th 2019, just before ICDAR 2019.

This special issue is a joint initiative of the board of editors of IJDAR and the Program Chairs of ICDAR 2019.

  • Andreas Dengel, Rafael Dueire Lins, Cheng-Lin Liu: Program Chairs of ICDAR 2019
  • Koichi Kise, Daniel Lopresti, Simone Marinai: Editors-in-Chief of IJDAR

Q&A

For any information about this initiative, please email one of the editors of this special issue or consult the FAQ that will be posted on the IJDAR website (https://www.springer.com/computer/image+processing/journal/10032).

Cheng-Lin Liu, ICDAR 2019 Program Co-Chair
( liucl@nlpr.ia.ac.cn )

CFP: International Workshop on Robust Reading (IWRR - repost)

3rd International Workshop on Robust Reading (IWRR)
ACCV 2018, Perth, Australia - December 2, 2018
IWRR Web Page

Important Dates:

September 28, 2018    Submission Deadline
  December 2, 2018    Workshop at ACCV

The 3rd IWRR workshop will be held in Perth, Australia, in conjunction with ACCV2018. The workshop aims at bringing together computer vision researchers and practitioners with an interest in reading systems that operate on images acquired in unconstrained conditions, such as scene images and video sequences, born-digital images, wearable camera and lifelog feeds, social media images, etc. The particular focus of the workshop is on the automatic extraction and interpretation of textual content in images, and applications that use textual information obtained automatically by such methods.

IWRR2018 invites the submission of original, previously unpublished work and welcomes re-submissions of improved versions of papers that have been rejected in the ACCV2018 conference reviewing process.

Workshop proceedings with accepted papers will be published along with the main conference proceedings by Springer in the Lecture Notes in Computer Science (LNCS) series.

The topics of interest include, among others:

  • Word spotting and end-to-end reading systems
  • Scene text based image retrieval
  • Joint modelling of textual and visual information
  • Text localisation, segmentation, and recognition in scene and born-digital images
  • Reading and tracking scene and/or overlaid text in video sequences
  • Robust reading applications (e.g. translation, reading text for the blind etc)
  • Performance evaluation and metrics

Visit the IWRR website for more information: http://www.cvc.uab.es/iwrr2018

Dimosthenis Karatzas, Workshop Co-Organizer
( dimos@cvc.uab.es )



IJDAR: Latest Issue (Vol. 21, Issue 3)

Special Issue on Deep Learning for Document Analysis and Recognition

Table of Contents

Click on the links to go directly to the Springer Link page for each article.

  1. Editors’ Message.

Cheng-Lin Liu, Gernot A. Fink, Venu Govindaraju, and Lianwen Jin

  1. Learning to detect, localize and recognize many text objects in document images from few examples.

Bastien Moysset, Christopher Kermorvant, and Christian Wolf

  1. Fully convolutional network with dilated convolutions for handwritten text line segmentation.

Guillaume Renton, ann Soullard, Clément Chatelain, Sébastien Adam, Christopher Kermorvant, and Thierry Paquet

  1. Integrating scattering feature maps with convolutional neural networks for Malayalam handwritten character recognition.

K. Manjusha, M. Anand Kumar, and K. P. Soman

  1. Attribute CNNs for word spotting in handwritten documents.

Sebastian Sudholt and Gernot A. Fink

  1. Fixed-sized representation learning from offline handwritten signatures of different sizes.

Luiz G. Hafemann, Luiz S. Oliveira, and Robert Sabourin

IJDAR Discount for IAPR Members (repost)

IAPR is pleased to announce a partnership agreement with Springer, the publisher of IJDAR, the International Journal on Document Analysis and Recognition. This new agreement will allow IAPR members to receive a subscription to the electronic version of IJDAR at a discount of nearly 50%. For additional details, see the links below:

Koichi Kise, Daniel Lopresti and Simone Marinai, IJDAR Editors-in-Chief
( kise@cs.osakafu-u.ac.jp, lopresti@cse.lehigh.edu, simone.marinai@unifi.it )



TC-11 maintains a colletion of datasets that can be found online in the TC-11 Datasets Repository.

If you have new datasets (e.g., from competitions) that you wish to share with the research community, please contact the TC-11 Dataset Curator (contact information is below).

Andreas Fischer (TC-11 Dataset Curator)
(andreas.fischer@unifr.ch)



IRISA/INSA Rennes (France): Research Engineer/PostDoc Position (3 Years - repost)

Analysis systems for serial sources in collections of historical image documents

Pdf version:
https://www-intuidoc.irisa.fr/en/postdoc-eurhisfirm

Important Dates

September 1, 2018 - August 31, 2021  Contract period

IRISA - Intuidoc
IRISA is a joint research center for Informatics, including Robotics and Image and Signal Processing. 800 people, 40 teams, explore the world of digital sciences to find applications in healthcare, ecology-environment, cyber-security, transportation, multimedia, and industry. INSA Rennes is one of the 8 trustees of IRISA.

The Intuidoc team (https://www.irisa.fr/intuidoc) conducts research on the topic of document image recognition. Since many years, the team proposes a system, called DMOS-PI method, for document structure analysis of documents. This DMOS-PI method is used for document recognition, or field extraction in archive documents, handwritten contents damaged documents (musical scores, archives, newspapers, letters, electronic schema, etc.).

EURHISFIRM project
EURHISFIRM European project aims at developing a research infrastructure to connect, collect, collate, align, and share reliable long-run company-level data for Europe to enable researchers, policymakers and other stakeholders to analyze, develop, and evaluate effective strategies to promote investment and economic growth. To achieve this goal, EURHISFIRM develops innovative tools to spark a “Big data” revolution in the historical social sciences and to open access to cultural heritage.

EURHISFIRM is a project funded by the European Commission within the Infrastructure Development Program of Horizon 2020. The first phase of the Infrastructure Development Program lasts for three years. It aims at developing an in-depth design study of the Research Infrastructure. After this phase, Development and Consolidation Phases follow if further applications will be successful. EURHISFIRM brings together eleven research institutions in economics, history, information technologies and data science from seven European countries.

Position to be filled

  • Position: Post-doctoral fellow / Research Engineer
  • Time commitment: Full-time
  • Duration of the contract: up to 36 months, starting as soon a possible
  • Supervisors: Bertrand Coüasnon and Aurélie Lemaitre
  • Indicative salary: Up to €36 000 gross annual salary (according to experience), with social security benefits
  • Location: IRISA – Rennes, France

Missions
The post-doctoral fellow / research engineer will be working on two tasks of EURHISFIRM workflow: the architecture of an adaptable system for document recognition, and the implementation of a generic structure layout extraction module.

The scientific challenge will be to extract information from various printed serial sources. Due to the large variety of those documents, a flexible and easy-to-adapt document recognition system is designed. For that purpose, the system will be based on a modeling of knowledge not only at the page level but also at the collection level in interaction with experts of the historical sources. Thus, redundancies between pages will be used to make the system more reliable and reduce manual corrections while obtaining a high recognition quality.

The system will we based on the DMOS-PI method which gives a framework for the analysis of collections of documents. It enables to share information from the collection between the pages, thanks to an iterative mechanism of analysis. This mechanism also makes it possible to integrate an asynchronous interaction between automatic analysis and human operators in order to limit the time of interaction by avoiding mutual waiting.

This modeling of the global analysis must be able to adapt to very different kinds of documents: from very structured documents, like stock exchange lists with redundancy and strong consistency between sequences of data, up to less structured documents, like yearbooks even if, also for them, the sequence from one year to another is important for improving the recognition quality.

The implementation of a generic structure extraction module will be based on the DMOS-PI method. It uses a grammatical language, EPF (Enhanced Position Formalism), to describe a general page layout, with perceptive vision mechanisms, and an iterative analysis. The system will also combine structural method with Deep Learning. For new collections, an adapted description of the document layout will be developed. This has to be done on a large range of structure levels: from very structured pages like table structures from stock exchange lists, up to a paragraph-oriented structures from yearbooks.

Applicant Requirements

  • PhD, Master degree or Engineering degree in computer science
  • Experience in document recognition, statistical analysis or deep learning.
  • Fluent English
  • Skills in grammars and languages and/or logical programming are nice-to-have.

For further information, please contact Bertrand Coüasnon (bertrand.couasnon@irisa.fr) and Aurélie Lemaitre (aurelie.lemaitre@irisa.fr). Applicants should send a curriculum-vitae with a list of publications and the names and email addresses of up to three references.

Bertrand Coüasnon, Director, Media and Interactions Department (IRISA)
( couasnon@irisa.fr )

Student Industrial Internship Opportunities (IAPR - repost)

IAPR’s Industrial Liaison Committee is pleased to announce the opening of its Company Internship Brokerage List.

The web page lists internship opportunities for students at different levels of education and specialism. We expect many additional internship opportunities to be listed here as the community becomes more aware of the site.

IAPR Company Internship Brokerage List:

http://homepages.inf.ed.ac.uk/rbf/IAPR/INDUSTRIAL

Bob Fisher, Chair, IAPR Industrial Liason Committee
( rbf@inf.ed.ac.uk )



Call for Contributions: To contribute news items, please send a short email to the editor, Richard Zanibbi ([1]). Contributions might include conference and workshop announcements/updates/reports, career opportunities, book reviews, or anything else of interest to the TC-11 community.

Subscription: This newsletter is sent to subscribers of the IAPR TC11 mailing list. To join the TC-11 mailing list, please click on this link: Join the TC-11 Mailing List. To manage your subscription, please visit the mailing list homepage: TC-11 Mailing List Homepage.