Paper Abstract and Keywords |
Presentation |
2022-05-13 10:15
Shape-preserving Style Transfer by Dense Pixel Correspondences between Input and Output Images and its Application to Gaze Estimation Daiki Mushiake, Norimichi Ukita (TTI) PRMU2022-2 |
Abstract |
(in Japanese) |
(See Japanese page) |
(in English) |
Gaze estimation has been applied to a variety of tasks, and many methods using deep learning have been proposed in recent years,but the annotation cost for training data is enormous. For this reason,a method has been proposed to generate training images by transforming automatically annotated CG images into real images using a style transfer.However, there is a problem that the iris and eyelids are shifted by pixel before and after the transformation to the training image, which adversely affects the training of the gaze estimator.The purpose of this study is to improve the accuracy of gaze estimator by suppressing the displacement of the iris and eyelids through a dense correspondence by pixels between the images before and after the transformation and by constraining the style transformer to keep the corresponding pixels on the same image coordinates. |
Keyword |
(in Japanese) |
(See Japanese page) |
(in English) |
Gaze Estimation / Style Transfer / / / / / / |
Reference Info. |
IEICE Tech. Rep., vol. 122, no. 13, PRMU2022-2, pp. 6-11, May 2022. |
Paper # |
PRMU2022-2 |
Date of Issue |
2022-05-05 (PRMU) |
ISSN |
Online edition: ISSN 2432-6380 |
Copyright and reproduction |
All rights are reserved and no part of this publication may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopy, recording, or any information storage and retrieval system, without permission in writing from the publisher. Notwithstanding, instructors are permitted to photocopy isolated articles for noncommercial classroom use without fee. (License No.: 10GA0019/12GB0052/13GB0056/17GB0034/18GB0034) |
Download PDF |
PRMU2022-2 |
Conference Information |
Committee |
PRMU IPSJ-CVIM |
Conference Date |
2022-05-12 - 2022-05-13 |
Place (in Japanese) |
(See Japanese page) |
Place (in English) |
Toyota Technological Institute |
Topics (in Japanese) |
(See Japanese page) |
Topics (in English) |
How to conduct research (post-graduation project for students) |
Paper Information |
Registration To |
PRMU |
Conference Code |
2022-05-PRMU-CVIM |
Language |
Japanese |
Title (in Japanese) |
(See Japanese page) |
Sub Title (in Japanese) |
(See Japanese page) |
Title (in English) |
Shape-preserving Style Transfer by Dense Pixel Correspondences between Input and Output Images and its Application to Gaze Estimation |
Sub Title (in English) |
|
Keyword(1) |
Gaze Estimation |
Keyword(2) |
Style Transfer |
Keyword(3) |
|
Keyword(4) |
|
Keyword(5) |
|
Keyword(6) |
|
Keyword(7) |
|
Keyword(8) |
|
1st Author's Name |
Daiki Mushiake |
1st Author's Affiliation |
Toyota Technological Institute (TTI) |
2nd Author's Name |
Norimichi Ukita |
2nd Author's Affiliation |
Toyota Technological Institute (TTI) |
3rd Author's Name |
|
3rd Author's Affiliation |
() |
4th Author's Name |
|
4th Author's Affiliation |
() |
5th Author's Name |
|
5th Author's Affiliation |
() |
6th Author's Name |
|
6th Author's Affiliation |
() |
7th Author's Name |
|
7th Author's Affiliation |
() |
8th Author's Name |
|
8th Author's Affiliation |
() |
9th Author's Name |
|
9th Author's Affiliation |
() |
10th Author's Name |
|
10th Author's Affiliation |
() |
11th Author's Name |
|
11th Author's Affiliation |
() |
12th Author's Name |
|
12th Author's Affiliation |
() |
13th Author's Name |
|
13th Author's Affiliation |
() |
14th Author's Name |
|
14th Author's Affiliation |
() |
15th Author's Name |
|
15th Author's Affiliation |
() |
16th Author's Name |
|
16th Author's Affiliation |
() |
17th Author's Name |
|
17th Author's Affiliation |
() |
18th Author's Name |
|
18th Author's Affiliation |
() |
19th Author's Name |
|
19th Author's Affiliation |
() |
20th Author's Name |
|
20th Author's Affiliation |
() |
Speaker |
Author-1 |
Date Time |
2022-05-13 10:15:00 |
Presentation Time |
15 minutes |
Registration for |
PRMU |
Paper # |
PRMU2022-2 |
Volume (vol) |
vol.122 |
Number (no) |
no.13 |
Page |
pp.6-11 |
#Pages |
6 |
Date of Issue |
2022-05-05 (PRMU) |