IW-FCV 2020

IW-FCV2020 was ended successfully.

Post-workshop proceedings have been published in Springer Link. (2020-5-4)

Awards were announced. (2020-2-27)

Updated registration fee can be found here. (2020-2-25)

Notice regarding to the Novel Coronavirus. (Update 2020-2-14)

Presentation instruction can be found here. (2020-2-3)

Detailed program is now available. (2020-1-29)

Online registration is now available. (2020-1-20)

Notice regarding to the Novel Coronavirus (SEE UPDATES BELOW - 2020-2-14)

As we all noticed the current outbreak of the new Corona Virus, the organizing committee would like to advise all participants of IW-FCV 2020 for the travel to Ibusuki, Japan as follows:

1) The immigration bureau of Japan does not permit the following people to enter Japan from Feb. 1st, 2020.

• Foreigners who have traveled to Hubei Province in China within 14 days before arriving in Japan,
• Foreigners who have Chinese passports issued by Hubei Province in China.

2) The organizing committee would like to ask for the participants, among whom visited Hubei Province during the one month and are doubtful about the virus, to cancel to attend the workshop.

3) The program committee will provide the opportunities to have an on-line (remote) presentation for those who cancel their trip to Japan due to Corona Virus.

4) When the participants travel through a big mass, such as an airport or train station, please pay much attention to individual health care. The following links are recommended for your information.

• http://www.cdc.go.kr/gallery.es?mid=a20503020000&bid=0003&b_list=9&act=view&list_no=144603&nPage=1&vlist_no_npage=1
• https://www.kantei.go.jp/jp/headline/kansensho/coronavirus.html
• https://www.who.int/emergencies/diseases/novel-coronavirus-2019

Since the current situation of novel coronavirus (COVID-19), we would like to provide some important notices, please take the time to read the following instructions carefully, and we kindly ask you to make your own decision.

1. If your condition matches at least one of the following situations, please consider to refrain from attendance at the workshop:

• 1-A. In case travel restrictions are announced by the government or your affiliated institutions such as universities.
• 1-B. In case of attendance becomes difficult due to severe circumstances, such as VISA problem or cancellation of air flights, etc.
• 1-C. In case you stayed or traveled nearby the regions infected by the novel coronavirus.

2. For authors who can not attend the workshop due to the above reasons, the workshop accepts a substitute presentation, remote presentation, and video presentation. Please consider the following options for your presentation:

• 2-A. Substitute oral/poster presentation.
  You will ask someone for substitute attendance. The organizing committee can not mediate between you and the substitutes.
• 2-B. Remote- or video- oral presentation.
  You consider having remote- or video- participation. The detail instructions for this option will be announced.
• 2-C. Substitute poster display.
  You will send us a PPT or PDF of your poster and prepare a short video (if possible). We plan to present them instead of you. The short video is not mandatory. Only the poster display is acceptable.
• 2-D. Cancel your presentation.
  You will cancel your presentation and withdraw your paper. When you choose this option, your registration fee will be fully refunded.

for registrants
3. For registrants who can not attend the workshop due to the above reasons, we accept cancellation of your registration with a full refund.

for registrants with a hotel, tour, and shuttle bus reservations
4. In case you booked hotel rooms, transportation, and tours at the same time of registration, you can cancel your reservations with a full refund.

for both authors and registrants
5. Refund policy
For authors who choose the option 2-D, the registration fee will be fully refunded. For the options 2-A to 2-C, at least one author should keep the registration. For registrants who canceled registrations, the registration fee will be fully refunded.

6. Instruction for cancellation
Please send us your request via the following google form by Sunday, Feb.16 23:59, in JST. After this due time, we do not guarantee to refund for your registration and reservations.

• Cancelation request form: announced by e-mail

a) Oral session :

• 15 minutes (10 minutes presentation and 5 minutes Q&A) presentation in an oral session
• paper type: full / short

b) Poster session :

• 1.5 hours presentation in a poster session
• paper type: full / short

c) Extended summary session :

• 7 minutes (5 minutes presentation and 2 minutes Q&A) presentation for the introduction of on-going projects in an oral session
• paper type: short

The organizing committee publishs the post-workshop proceedings of IW-FCV2020 as a Springer Communications in Computer and Information Science (CCIS) series after the workshop. The post-workshop proceedigns will include only high quality papers selected by reviewers.

The conditions under which a paper is subject to peer review for inclusion in the post-workshop proceedings are as follows:
1.The paper must be a full paper (12-15+ pages) which contains original research that has neither been published nor submitted for publication elsewhere.
2.When submitting a paper, the authors indicates his/her intension to include the paper in the post-workshop proceedings.

In addition, in order for a paper to be included in the post-workshop proceedings, it must:.
3. be presented at the workshop by at least one author.
4. When submitting the camera-ready version, the authors complete and sign a Consent-to-Publish form to transfer the copyright for the paper to publisher.

The post-workshop proceedings will be included in Springer Link Digital Library. And workshop attendees have free access for four weeks after the post-workshop proceedings are published.

Authors should consult Springer's authors' guidelines and use their proceedings templates for the preparation of their papers. Springer encourages authors to include their ORCIDs in their papers. Please kindly refer to the editorial policies webpage and note the importance of this.

IW-FCV 2020 welcomes submissions for DEMO presentations.

Your demo presentations will be shown in the workshop program (Web, and booklet) and be presented in the same room as the poster presentations.

A workshop paper is NOT mandatory for the demo presentation, but you can submit an article to appear in the workshop proceedings (USB stick).
The demo papers will not appear in the post-workshop proceedings.

For submitting the demo presentation, please send the program chair ([email protected]) an e-mail containing the following items.

The due date of demo submission is Jan. 31, 2020.

We can provide a board for posters, one or two tables, and a power supply.
Please note the type of electric outlet is Japanese (Type-A).
We can not provide any converter for the outlet type.

Each oral presentation is allocated 15 minutes: 10 minutes for presentation + 5 minutes for Q&A and setup. And each extended-summary presentation is allocated 5 minutes for presentation + 2 minutes for Q&A and setup.

At the workshop, 10 minutes before your session, you must:

The LCD projector has both VGA and HDMI connections. While the projector is compatible with wide-size (16:9 or 16:10) presentation slides, the size (aspect ratio) of the screen is standard-size (4:3). So, be sure that a wide-size presentation will be reduced in size to fit the screen width.

The speed of the internet connection will be limited. We do not recommend you to use video-streaming or online-demonstration in your presentation.

The poster panels can accommodate posters of A0 portrait size. Please use sufficiently large fonts for readability. We do not specify any poster templates and can not provide poster-printing services.

The location of your poster will be identified by the paper number from the program. Materials to attach your poster to the board (before 9 am of your day of presentation) will be provided. Each poster is to be presented for a full day.

Prof. Changick Kim, KAIST (Korea)

• Web site
• Title - Domain adaptation and its application to object detection
Abstract

During the last decade, the field of Computer vision has made huge progress. This progress is mostly due to the powerful effectiveness of Convolutional Neural Networks (CNNs). CNNs allow for very precise predictions if trained with high-quality annotated training data. In a classification setting for instance, you would often use one of the standard network architectures out there (ResNet, VGG etc.) and train it using your dataset. This would likely lead to very good performance. However, supervised learning-based methods perform poorly when the train data and the test data have different distributions. To address the issue, domain adaptation transfers knowledge from the label-sufficient domain (source domain) to the label-scarce domain (target domain). In this talk, I will address the basic concepts of transfer learning and domain adaptation, and talk about multiple approaches to domain adaptation. In addition, recent domain adaptation algorithms applied to object detection will be explained. This talk will conclude with arguments that there is still a lot to be done before domain-adaptive learning systems become practical tools.

Dr. Yoshitaka Ushiku, OMRON SINIC X Corporation (Japan)

• Web site
• Title - Deep Learning for Natural Language Processing and Computer Vision
Abstract

In 2012, computer vision (CV) was benefited from deep learning, resulting in dramatic improvements in image recognition performance. Nowadays, various applications, including detection, segmentation, and 3D vision, are widely studied utilizing state-of-the-art learning methods. In addition to the applications of machine learning (ML), CV area has also been a testbed for fundamental ML technologies, such as convolutional neural networks and transfer learning. Unsupervised domain adaptation (UDA) is a task to transfer knowledge obtained from a source domain to another target domain. The difficulty is that the target samples are not associated with teaching labels, while all data in the source domain are labeled. In the former part of this talk, I will introduce such cutting-edge tasks in CV and ML.

Besides, with the development of recent machine learning techniques and the commoditization of them, the barriers to import the latest technologies from other areas are dramatically narrowed. Visual captioning, which is a task to generate fluent captions to describe the input multimedia datum, is proposed as such integration of ideas from different fields, namely CV and natural language processing (NLP). Caption generation and its variational problems are energetically addressed from both CV and NLP researchers. Thus, vision and language, which is an intersection area of CV and NLP, is spreading widely. It also includes visual question answering, a task to answer questions with a related photo that visually tells the correct answer. Another task that attracts attention from CV, NLP, and even ML researchers is visual dialog, which is a task that agents attempt to talk to each other concerning visual information. For example, salesclerks use the visual information of their merchandise to sell them to customers. In the latter part of this talk, I will introduce some challenging tasks in CV and some additional topics in vision and language.

Prof. Md Atiqur Rahman Ahad, University of Dhaka (Bangladesh)

• Web site
• Title - Activity & Gait Recognition: Applications in Healthcare & Challenges Ahead
Abstract

Human activity recognitions in both video and sensor domains are very important for different applications. Gait or walking pattern is an important biometric cue for human identification and verification. Understanding of activity or gait can be useful for various applications related to healthcare, elderly support, security, gender identification, etc. These are progressing and various challenges are ahead. There is huge progress recently on vision-based activity and gait analyses – because of the availability of GPUs and deep learning-based methods. Sensor-based gait recognition approaches are explored as well in different areas. However, the challenges are complex and diversified, and hence, the progresses in these domains are still constrained by several genuine challenges. In this speech, some of these scopes and challenges will be highlighted. The talk will highlight our recent competition, held in the IEEE 12th International Conference on Biometrics (ICB), 2019, where 18 teams registered from 14 countries. Finally, 67 solutions were submitted from 10 teams – for age and gender estimation. Occlusion and other challenges will be presented.