Sparse Single Photon Reconstruction through Masked Autoencoder

Luke McEvoy; Daniel Tafone; Yong Meng Sua; Yuping Huang

Sparse Single Photon Reconstruction through Masked Autoencoder

Luke McEvoy, Daniel Tafone, Yong Meng Sua, Yuping Huang

Department of Physics

Stevens Institute of Technology

Research output: Contribution to conference › Paper › peer-review

Abstract

Leveraging masked autoencoders, we achieve 4 times classification accuracy than that of Vision Transformers, and 2 times better classification accuracy than Convolutional Neural Networks on 50% masked LiDAR images acquired with sparse photon detections.

Original language	English
State	Published - 2023
Event	2023 Laser Science, LS 2023 - Tacoma, United States Duration: 9 Oct 2023 → 12 Oct 2023

Conference

Conference	2023 Laser Science, LS 2023
Country/Territory	United States
City	Tacoma
Period	9/10/23 → 12/10/23

Cite this

@conference{1e77490d90e24528b43f64f676a318e7,

title = "Sparse Single Photon Reconstruction through Masked Autoencoder",

abstract = "Leveraging masked autoencoders, we achieve 4 times classification accuracy than that of Vision Transformers, and 2 times better classification accuracy than Convolutional Neural Networks on 50\% masked LiDAR images acquired with sparse photon detections.",

author = "Luke McEvoy and Daniel Tafone and Sua, \{Yong Meng\} and Yuping Huang",

note = "Publisher Copyright: {\textcopyright} Optica Publishing Group 2023, {\textcopyright} 2023 The Authors.; 2023 Laser Science, LS 2023 ; Conference date: 09-10-2023 Through 12-10-2023",

year = "2023",

language = "English",

}

TY - CONF

T1 - Sparse Single Photon Reconstruction through Masked Autoencoder

AU - McEvoy, Luke

AU - Tafone, Daniel

AU - Sua, Yong Meng

AU - Huang, Yuping

PY - 2023

Y1 - 2023

N2 - Leveraging masked autoencoders, we achieve 4 times classification accuracy than that of Vision Transformers, and 2 times better classification accuracy than Convolutional Neural Networks on 50% masked LiDAR images acquired with sparse photon detections.

AB - Leveraging masked autoencoders, we achieve 4 times classification accuracy than that of Vision Transformers, and 2 times better classification accuracy than Convolutional Neural Networks on 50% masked LiDAR images acquired with sparse photon detections.

UR - https://www.scopus.com/pages/publications/85215510671

UR - https://www.scopus.com/inward/citedby.url?scp=85215510671&partnerID=8YFLogxK

M3 - Paper

AN - SCOPUS:85215510671

T2 - 2023 Laser Science, LS 2023

Y2 - 9 October 2023 through 12 October 2023

ER -

Sparse Single Photon Reconstruction through Masked Autoencoder

Abstract

Conference

Other files and links

Cite this