TY - JOUR
T1 - Target recognition in ghost imaging from traditional to advance; a brief review
AU - Abbas, Ayesha
AU - Mu, Jianbang
AU - Mengyue, Zhang
AU - Cao, Jie
AU - Zhang, Xiaonan
N1 - Publisher Copyright:
© 2025
PY - 2025/6
Y1 - 2025/6
N2 - Ghost Imaging (GI) has emerged as groundbreaking technique in image processing, offering unique advantages such as non-locality, simplified optical setups, high detection efficiency and robust performance in scattering media. Over the years, GI has evolved from fundamental research to real time applications, driven by advancements in methodologies and computational techniques. However, traditional optical approaches for target recognition often encounter limitations, particularly in achieving high accuracy and reliability under complex conditions. To overcome these challenges, the integration of deep learning (DL) and machine learning (ML) into GI systems presents a promising solutions, enabling enhanced performances with reduced constraints. This review provides a comprehensive analysis of target recognition methodologies in GI, comparing traditional optical techniques with cutting-edge DL and ML- based approaches. Furthermore, we discuss the key challenges faced by these methods and propose future directions to advance this innovative and interdisciplinary field, paving the way for more precise and scalable solutions in practical applications.
AB - Ghost Imaging (GI) has emerged as groundbreaking technique in image processing, offering unique advantages such as non-locality, simplified optical setups, high detection efficiency and robust performance in scattering media. Over the years, GI has evolved from fundamental research to real time applications, driven by advancements in methodologies and computational techniques. However, traditional optical approaches for target recognition often encounter limitations, particularly in achieving high accuracy and reliability under complex conditions. To overcome these challenges, the integration of deep learning (DL) and machine learning (ML) into GI systems presents a promising solutions, enabling enhanced performances with reduced constraints. This review provides a comprehensive analysis of target recognition methodologies in GI, comparing traditional optical techniques with cutting-edge DL and ML- based approaches. Furthermore, we discuss the key challenges faced by these methods and propose future directions to advance this innovative and interdisciplinary field, paving the way for more precise and scalable solutions in practical applications.
KW - Deep learning
KW - Ghost Imaging
KW - Target recognition
KW - Traditional methods of target recognition
UR - http://www.scopus.com/inward/record.url?scp=85214517349&partnerID=8YFLogxK
U2 - 10.1016/j.optlastec.2025.112450
DO - 10.1016/j.optlastec.2025.112450
M3 - Review article
AN - SCOPUS:85214517349
SN - 0030-3992
VL - 184
JO - Optics and Laser Technology
JF - Optics and Laser Technology
M1 - 112450
ER -