Information: Join today and make your research activities more affordable! Technical workshop participation fees and annual registration fees are available at member rates.
Notice: [Important] Announcement of Changes to Registration Fee Payment and Manuscript Upload Procedures for IEICE Technical Meetings
IEICE Technical Committee Submission System
Conference Paper's Information		 Online Proceedings
[Sign in]
Tech. Rep. Archives
 Go Top Page Go Previous   [Japanese] / [English] 

Paper Abstract and Keywords
Presentation 2023-05-19 15:25
Prompt Learning for Object Detection with Vision-Language Model
Mariko Tomariguchi (OKI) PRMU2023-12
Abstract (in Japanese) (See Japanese page) 
(in English) The two-stage object detection models crop features in the regions where objects are most likely to be to classify the objects. In this work, we investigate the influence of the surrounding information on the objects on classifying objects and improve the prompt learning method for object detection using Vision-Language models. We learn the learnable vectors correspond to input prompts to CLIP with augmented data to create prompts with and without surroundings information. Then, we train the object detection model substituting the calculation of the classification score for the language embedding obtained from passing the learned prompts through the CLIP language encoder. Our method achieves 20.3 %$mathrm{AP}$ on the LVIS dataset with prompts including surroundings, and 21.6 %$mathrm{AP}$ with prompts not including surroundings. In particular, 27.9 % mathrm{AP}_f$ and 29.1 % $mathrm{AP}_f$ are achieved in the LVIS frequency class, respectively.
Keyword (in Japanese) (See Japanese page) 
(in English) deep learning / object detection / Mask R-CNN / prompt learning / CLIP / / /  
Reference Info. IEICE Tech. Rep., vol. 123, no. 30, PRMU2023-12, pp. 62-67, May 2023.
Paper # PRMU2023-12 
Date of Issue 2023-05-11 (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 PRMU2023-12

Conference Information
Committee PRMU IPSJ-CVIM  
Conference Date 2023-05-18 - 2023-05-19 
Place (in Japanese) (See Japanese page) 
Place (in English)  
Topics (in Japanese) (See Japanese page) 
Topics (in English)  
Paper Information
Registration To PRMU 
Conference Code 2023-05-PRMU-CVIM 
Language Japanese 
Title (in Japanese) (See Japanese page) 
Sub Title (in Japanese) (See Japanese page) 
Title (in English) Prompt Learning for Object Detection with Vision-Language Model 
Sub Title (in English)  
Keyword(1) deep learning  
Keyword(2) object detection  
Keyword(3) Mask R-CNN  
Keyword(4) prompt learning  
Keyword(5) CLIP  
Keyword(6)  
Keyword(7)  
Keyword(8)  
1st Author's Name Mariko Tomariguchi  
1st Author's Affiliation Oki Electric Industry Co., Ltd. (OKI)
2nd Author's Name  
2nd Author's Affiliation ()
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 ()
21st Author's Name  
21st Author's Affiliation ()
22nd Author's Name  
22nd Author's Affiliation ()
23rd Author's Name  
23rd Author's Affiliation ()
24th Author's Name  
24th Author's Affiliation ()
25th Author's Name  
25th Author's Affiliation ()
26th Author's Name / /
26th Author's Affiliation ()
()
27th Author's Name / /
27th Author's Affiliation ()
()
28th Author's Name / /
28th Author's Affiliation ()
()
29th Author's Name / /
29th Author's Affiliation ()
()
30th Author's Name / /
30th Author's Affiliation ()
()
31st Author's Name / /
31st Author's Affiliation ()
()
32nd Author's Name / /
32nd Author's Affiliation ()
()
33rd Author's Name / /
33rd Author's Affiliation ()
()
34th Author's Name / /
34th Author's Affiliation ()
()
35th Author's Name / /
35th Author's Affiliation ()
()
36th Author's Name / /
36th Author's Affiliation ()
()
Speaker Author-1 
Date Time 2023-05-19 15:25:00 
Presentation Time 15 minutes 
Registration for PRMU 
Paper # PRMU2023-12 
Volume (vol) vol.123 
Number (no) no.30 
Page pp.62-67 
#Pages 6 
Date of Issue 2023-05-11 (PRMU) 

[Return to Top Page]

[Return to IEICE Web Page]
The Institute of Electronics, Information and Communication Engineers (IEICE), Japan