PS-YOLO-seg: A Lightweight Instance Segmentation Method for Lithium Mineral Microscopic Images Based on Improved YOLOv12-seg

Microscopic image automatic recognition is a core technology for mineral composition analysis and plays a crucial role in advancing the intelligent development of smart mining systems. To overcome the limitations of traditional lithium ore analysis and meet the challenges of deployment on edge devic...

Full description

Saved in:
Bibliographic Details
Main Authors: Zeyang Qiu, Xueyu Huang, Zhicheng Deng, Xiangyu Xu, Zhenzhong Qiu
Format: Article
Language:English
Published: MDPI AG 2025-07-01
Series:Journal of Imaging
Subjects:
Online Access:https://www.mdpi.com/2313-433X/11/7/230
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1849732851844513792
author Zeyang Qiu
Xueyu Huang
Zhicheng Deng
Xiangyu Xu
Zhenzhong Qiu
author_facet Zeyang Qiu
Xueyu Huang
Zhicheng Deng
Xiangyu Xu
Zhenzhong Qiu
author_sort Zeyang Qiu
collection DOAJ
description Microscopic image automatic recognition is a core technology for mineral composition analysis and plays a crucial role in advancing the intelligent development of smart mining systems. To overcome the limitations of traditional lithium ore analysis and meet the challenges of deployment on edge devices, we propose PS-YOLO-seg, a lightweight segmentation model specifically designed for lithium mineral analysis under visible light microscopy. The network is compressed by adjusting the width factor to reduce global channel redundancy. A PSConv-based downsampling strategy enhances the network’s ability to capture dim mineral textures under microscopic conditions. In addition, the improved C3k2-PS module strengthens feature extraction, while the streamlined Segment-Efficient head minimizes redundant computation, further reducing the overall model complexity. PS-YOLO-seg achieves a slightly improved segmentation performance compared to the baseline YOLOv12n model on a self-constructed lithium ore microscopic dataset, while reducing FLOPs by 20%, parameter count by 33%, and model size by 32%. Additionally, it achieves a faster inference speed, highlighting its potential for practical deployment. This work demonstrates how architectural optimization and targeted enhancements can significantly improve instance segmentation performance while maintaining speed and compactness, offering strong potential for real-time deployment in industrial settings and edge computing scenarios.
format Article
id doaj-art-ebd90191f35b422cae6977bfee06124f
institution DOAJ
issn 2313-433X
language English
publishDate 2025-07-01
publisher MDPI AG
record_format Article
series Journal of Imaging
spelling doaj-art-ebd90191f35b422cae6977bfee06124f2025-08-20T03:08:12ZengMDPI AGJournal of Imaging2313-433X2025-07-0111723010.3390/jimaging11070230PS-YOLO-seg: A Lightweight Instance Segmentation Method for Lithium Mineral Microscopic Images Based on Improved YOLOv12-segZeyang Qiu0Xueyu Huang1Zhicheng Deng2Xiangyu Xu3Zhenzhong Qiu4Yichun Lithium New Energy Industry Research Institute, Jiangxi University of Science and Technology, Yichun 336000, ChinaYichun Lithium New Energy Industry Research Institute, Jiangxi University of Science and Technology, Yichun 336000, ChinaSchool of Software Engineering, Jiangxi University of Science and Technology, Nanchang 330013, ChinaSchool of Software Engineering, Jiangxi University of Science and Technology, Nanchang 330013, ChinaJiangxi Yongxing Special Steel New Energy Technology Co., Ltd., Yichun 336300, ChinaMicroscopic image automatic recognition is a core technology for mineral composition analysis and plays a crucial role in advancing the intelligent development of smart mining systems. To overcome the limitations of traditional lithium ore analysis and meet the challenges of deployment on edge devices, we propose PS-YOLO-seg, a lightweight segmentation model specifically designed for lithium mineral analysis under visible light microscopy. The network is compressed by adjusting the width factor to reduce global channel redundancy. A PSConv-based downsampling strategy enhances the network’s ability to capture dim mineral textures under microscopic conditions. In addition, the improved C3k2-PS module strengthens feature extraction, while the streamlined Segment-Efficient head minimizes redundant computation, further reducing the overall model complexity. PS-YOLO-seg achieves a slightly improved segmentation performance compared to the baseline YOLOv12n model on a self-constructed lithium ore microscopic dataset, while reducing FLOPs by 20%, parameter count by 33%, and model size by 32%. Additionally, it achieves a faster inference speed, highlighting its potential for practical deployment. This work demonstrates how architectural optimization and targeted enhancements can significantly improve instance segmentation performance while maintaining speed and compactness, offering strong potential for real-time deployment in industrial settings and edge computing scenarios.https://www.mdpi.com/2313-433X/11/7/230instance segmentationlepidoliteYOLOv12-seglightweight networkedge deploymentmicroscopic imaging
spellingShingle Zeyang Qiu
Xueyu Huang
Zhicheng Deng
Xiangyu Xu
Zhenzhong Qiu
PS-YOLO-seg: A Lightweight Instance Segmentation Method for Lithium Mineral Microscopic Images Based on Improved YOLOv12-seg
Journal of Imaging
instance segmentation
lepidolite
YOLOv12-seg
lightweight network
edge deployment
microscopic imaging
title PS-YOLO-seg: A Lightweight Instance Segmentation Method for Lithium Mineral Microscopic Images Based on Improved YOLOv12-seg
title_full PS-YOLO-seg: A Lightweight Instance Segmentation Method for Lithium Mineral Microscopic Images Based on Improved YOLOv12-seg
title_fullStr PS-YOLO-seg: A Lightweight Instance Segmentation Method for Lithium Mineral Microscopic Images Based on Improved YOLOv12-seg
title_full_unstemmed PS-YOLO-seg: A Lightweight Instance Segmentation Method for Lithium Mineral Microscopic Images Based on Improved YOLOv12-seg
title_short PS-YOLO-seg: A Lightweight Instance Segmentation Method for Lithium Mineral Microscopic Images Based on Improved YOLOv12-seg
title_sort ps yolo seg a lightweight instance segmentation method for lithium mineral microscopic images based on improved yolov12 seg
topic instance segmentation
lepidolite
YOLOv12-seg
lightweight network
edge deployment
microscopic imaging
url https://www.mdpi.com/2313-433X/11/7/230
work_keys_str_mv AT zeyangqiu psyolosegalightweightinstancesegmentationmethodforlithiummineralmicroscopicimagesbasedonimprovedyolov12seg
AT xueyuhuang psyolosegalightweightinstancesegmentationmethodforlithiummineralmicroscopicimagesbasedonimprovedyolov12seg
AT zhichengdeng psyolosegalightweightinstancesegmentationmethodforlithiummineralmicroscopicimagesbasedonimprovedyolov12seg
AT xiangyuxu psyolosegalightweightinstancesegmentationmethodforlithiummineralmicroscopicimagesbasedonimprovedyolov12seg
AT zhenzhongqiu psyolosegalightweightinstancesegmentationmethodforlithiummineralmicroscopicimagesbasedonimprovedyolov12seg