Proposal of an object recognition method capable of detecting and clustering unknown objects

Yu Horiuchi; Shuichi Akizuki; Manabu Hashimoto; Takuma Mori; Kenji Nishimiya

doi:10.1117/12.3018711

2 May 2024 Proposal of an object recognition method capable of detecting and clustering unknown objects

Yu Horiuchi, Shuichi Akizuki, Manabu Hashimoto, Takuma Mori, Kenji Nishimiya

Author Affiliations +

Proceedings Volume 13164, International Workshop on Advanced Imaging Technology (IWAIT) 2024; 131641X (2024) https://doi.org/10.1117/12.3018711
Event: International Workshop on Advanced Imaging Technology (IWAIT) 2024, 2024, Langkawi, Malaysia

Abstract

The proposed object recognition method can handle both known and unknown objects in an image. For known objects, the proposed method can classify objects as accurately as conventional methods do. If the image contains unlearned objects, the method detects the unknown objects. The method even has the ability to classify the objects as multiple classes. During the evaluation phase of the study, we compared the proposed method with conventional object detection methods using mean average precision (mAP). The proposed method’s known object classification accuracy was 47.1, while that of the conventional method was 46.8. This indicates that the proposed method and the compared method have the same accuracy in classifying known objects. The accuracy for unknown objects was 23.6, indicating that unknown objects can also be classified. Thus, it can be concluded that the proposed method can classify known objects with high accuracy and also deal with unknown objects.

(2024) Published by SPIE. Downloading of the abstract is permitted for personal use only.

Citation Download Citation

Yu Horiuchi, Shuichi Akizuki, Manabu Hashimoto, Takuma Mori, and Kenji Nishimiya "Proposal of an object recognition method capable of detecting and clustering unknown objects", Proc. SPIE 13164, International Workshop on Advanced Imaging Technology (IWAIT) 2024, 131641X (2 May 2024); https://doi.org/10.1117/12.3018711

ACCESS THE FULL ARTICLE

INSTITUTIONAL
Select your institution to access the SPIE Digital Library.

SELECT YOUR INSTITUTION

PERSONAL
Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.

PERSONAL SIGN IN

No SPIE Account? Create one

;

PURCHASE THIS CONTENT

SUBSCRIBE TO DIGITAL LIBRARY

50 downloads per 1-year subscription

Members: $195

Non-members: $335 ADD TO CART

25 downloads per 1 - year subscription

Members: $145

Non-members: $250 ADD TO CART

PURCHASE SINGLE ARTICLE