Latent Space Search-Based Adaptive Template Generation for Enhanced Object Detection in Bin-Picking Applications

Template matching is a common approach in bin-picking tasks. However, it often struggles in complex environments, such as those with different object poses, various background appearances, and varying lighting conditions, due to the limited feature representation of a single template. Additionally,...

Full description

Saved in:
Bibliographic Details
Main Authors: Songtao Liu, Yaonan Zhu, Tadayoshi Aoyama, Masayuki Nakaya, Yasuhisa Hasegawa
Format: Article
Language:English
Published: MDPI AG 2024-09-01
Series:Sensors
Subjects:
Online Access:https://www.mdpi.com/1424-8220/24/18/6050
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1850259480283971584
author Songtao Liu
Yaonan Zhu
Tadayoshi Aoyama
Masayuki Nakaya
Yasuhisa Hasegawa
author_facet Songtao Liu
Yaonan Zhu
Tadayoshi Aoyama
Masayuki Nakaya
Yasuhisa Hasegawa
author_sort Songtao Liu
collection DOAJ
description Template matching is a common approach in bin-picking tasks. However, it often struggles in complex environments, such as those with different object poses, various background appearances, and varying lighting conditions, due to the limited feature representation of a single template. Additionally, during the bin-picking process, the template needs to be frequently updated to maintain detection performance, and finding an adaptive template from a vast dataset poses another challenge. To address these challenges, we propose a novel template searching method in a latent space trained by a Variational Auto-Encoder (VAE), which generates an adaptive template dynamically based on the current environment. The proposed method was evaluated experimentally under various conditions, and in all scenarios, it successfully completed the tasks, demonstrating its effectiveness and robustness for bin-picking applications. Furthermore, we integrated our proposed method with YOLO, and the experimental results indicate that our method effectively improves YOLO’s detection performance.
format Article
id doaj-art-ef3eeddcb5214bff8ecd449163b71d35
institution OA Journals
issn 1424-8220
language English
publishDate 2024-09-01
publisher MDPI AG
record_format Article
series Sensors
spelling doaj-art-ef3eeddcb5214bff8ecd449163b71d352025-08-20T01:55:51ZengMDPI AGSensors1424-82202024-09-012418605010.3390/s24186050Latent Space Search-Based Adaptive Template Generation for Enhanced Object Detection in Bin-Picking ApplicationsSongtao Liu0Yaonan Zhu1Tadayoshi Aoyama2Masayuki Nakaya3Yasuhisa Hasegawa4Department of Micro-Nano Mechanical Science and Engineering, Nagoya University, Nagoya 464-8601, JapanThe School of Engineering, The University of Tokyo, 7-3-1 Hongo, Bunkyo-ku, Tokyo 113-8656, JapanDepartment of Micro-Nano Mechanical Science and Engineering, Nagoya University, Nagoya 464-8601, JapanRobot Division, System Department, NACHI-FUJIKOSHI CORP., Toyama 930-8511, JapanDepartment of Micro-Nano Mechanical Science and Engineering, Nagoya University, Nagoya 464-8601, JapanTemplate matching is a common approach in bin-picking tasks. However, it often struggles in complex environments, such as those with different object poses, various background appearances, and varying lighting conditions, due to the limited feature representation of a single template. Additionally, during the bin-picking process, the template needs to be frequently updated to maintain detection performance, and finding an adaptive template from a vast dataset poses another challenge. To address these challenges, we propose a novel template searching method in a latent space trained by a Variational Auto-Encoder (VAE), which generates an adaptive template dynamically based on the current environment. The proposed method was evaluated experimentally under various conditions, and in all scenarios, it successfully completed the tasks, demonstrating its effectiveness and robustness for bin-picking applications. Furthermore, we integrated our proposed method with YOLO, and the experimental results indicate that our method effectively improves YOLO’s detection performance.https://www.mdpi.com/1424-8220/24/18/6050template matchingtemplate searchingtemplate generationbin picking
spellingShingle Songtao Liu
Yaonan Zhu
Tadayoshi Aoyama
Masayuki Nakaya
Yasuhisa Hasegawa
Latent Space Search-Based Adaptive Template Generation for Enhanced Object Detection in Bin-Picking Applications
Sensors
template matching
template searching
template generation
bin picking
title Latent Space Search-Based Adaptive Template Generation for Enhanced Object Detection in Bin-Picking Applications
title_full Latent Space Search-Based Adaptive Template Generation for Enhanced Object Detection in Bin-Picking Applications
title_fullStr Latent Space Search-Based Adaptive Template Generation for Enhanced Object Detection in Bin-Picking Applications
title_full_unstemmed Latent Space Search-Based Adaptive Template Generation for Enhanced Object Detection in Bin-Picking Applications
title_short Latent Space Search-Based Adaptive Template Generation for Enhanced Object Detection in Bin-Picking Applications
title_sort latent space search based adaptive template generation for enhanced object detection in bin picking applications
topic template matching
template searching
template generation
bin picking
url https://www.mdpi.com/1424-8220/24/18/6050
work_keys_str_mv AT songtaoliu latentspacesearchbasedadaptivetemplategenerationforenhancedobjectdetectioninbinpickingapplications
AT yaonanzhu latentspacesearchbasedadaptivetemplategenerationforenhancedobjectdetectioninbinpickingapplications
AT tadayoshiaoyama latentspacesearchbasedadaptivetemplategenerationforenhancedobjectdetectioninbinpickingapplications
AT masayukinakaya latentspacesearchbasedadaptivetemplategenerationforenhancedobjectdetectioninbinpickingapplications
AT yasuhisahasegawa latentspacesearchbasedadaptivetemplategenerationforenhancedobjectdetectioninbinpickingapplications