Deep Learning Techniques for Gamma Ray Interaction Location Estimation in Monolithic Scintillation Crystal Detectors

Li Tao, Xin Li, Lars R. Furenlid, Craig S. Levin

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

Monolithic scintillation crystals have been widely studied as detectors for PET systems. In this work, we explore deep learning techniques using the information from mean detector response functions (MDRFs) as a new, potentially faster and more accurate method to estimate gamma ray interaction location in monolithic scintillation crystal detectors. Compared with searching based methods, deep learning techniques do not require recording all the MDRF information once the prediction networks are trained. This could reduce the memory cost by a factor of up to 100. We have designed four different neural networks to estimate/predict gamma ray interaction location given the MDRF data. The networks are trained with calibrated MDRFs at 252×252 locations in a 50.8×50.8 mm2 area, with each MDRF data point containing the information from 20 SiPM channels. Our first network consists of only fully connected (FC) layers, with a final regression layer that directly predicts the values for x and y location. This network is trained to eventually have a loss of 4.8 (L2 loss in units of mm2). The second network is also consisted of only FC layers, but with a final classification layer that classifies the x and y location into 252 classes. This network is trained to have a final prediction accuracy of 15%. The third network is designed as a convolutional neural network (CNN) with a final classification layer. This network is trained to have a final prediction accuracy of 20%. With the last network, we perform "sectional training" by first coarsely separating the entire crystal into 12×12 sub-areas. Then for each sub-area, we do fine training and classify the x and y location into the correct class. This network is trained to have a final prediction accuracy of 90%. We test the trained networks with a 5 slit image. The RMS prediction errors for the four networks are 2.6 mm (FC regression network), 2.2 mm (FC classification network), 2.1 mm (CNN network) and 2.0 mm ("sectional training" network). We can see that the CNN network and "sectional training" network can achieve lower prediction error. In comparison with searching based methods, deep learning based estimation methods do not need to keep record of all the MDRF information, which consists of a total of 1270080 parameters, once the networks are trained. This has reduced the memory cost by a factor of 10 - 100 depending on the network structure.

Original languageEnglish (US)
Title of host publication2018 IEEE Nuclear Science Symposium and Medical Imaging Conference, NSS/MIC 2018 - Proceedings
PublisherInstitute of Electrical and Electronics Engineers Inc.
ISBN (Electronic)9781538684948
DOIs
StatePublished - Nov 2018
Event2018 IEEE Nuclear Science Symposium and Medical Imaging Conference, NSS/MIC 2018 - Sydney, Australia
Duration: Nov 10 2018Nov 17 2018

Publication series

Name2018 IEEE Nuclear Science Symposium and Medical Imaging Conference, NSS/MIC 2018 - Proceedings

Conference

Conference2018 IEEE Nuclear Science Symposium and Medical Imaging Conference, NSS/MIC 2018
CountryAustralia
CitySydney
Period11/10/1811/17/18

Fingerprint

Gamma Rays
learning
scintillation
Learning
gamma rays
detectors
crystals
interactions
Costs and Cost Analysis
education
predictions
regression analysis

Keywords

  • deep learning
  • gamma ray interaction location prediction
  • mean detector response function
  • Monolithic scintillation detector
  • neural network

ASJC Scopus subject areas

  • Radiology Nuclear Medicine and imaging
  • Nuclear and High Energy Physics

Cite this

Tao, L., Li, X., Furenlid, L. R., & Levin, C. S. (2018). Deep Learning Techniques for Gamma Ray Interaction Location Estimation in Monolithic Scintillation Crystal Detectors. In 2018 IEEE Nuclear Science Symposium and Medical Imaging Conference, NSS/MIC 2018 - Proceedings [8824365] (2018 IEEE Nuclear Science Symposium and Medical Imaging Conference, NSS/MIC 2018 - Proceedings). Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/NSSMIC.2018.8824365

Deep Learning Techniques for Gamma Ray Interaction Location Estimation in Monolithic Scintillation Crystal Detectors. / Tao, Li; Li, Xin; Furenlid, Lars R.; Levin, Craig S.

2018 IEEE Nuclear Science Symposium and Medical Imaging Conference, NSS/MIC 2018 - Proceedings. Institute of Electrical and Electronics Engineers Inc., 2018. 8824365 (2018 IEEE Nuclear Science Symposium and Medical Imaging Conference, NSS/MIC 2018 - Proceedings).

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Tao, L, Li, X, Furenlid, LR & Levin, CS 2018, Deep Learning Techniques for Gamma Ray Interaction Location Estimation in Monolithic Scintillation Crystal Detectors. in 2018 IEEE Nuclear Science Symposium and Medical Imaging Conference, NSS/MIC 2018 - Proceedings., 8824365, 2018 IEEE Nuclear Science Symposium and Medical Imaging Conference, NSS/MIC 2018 - Proceedings, Institute of Electrical and Electronics Engineers Inc., 2018 IEEE Nuclear Science Symposium and Medical Imaging Conference, NSS/MIC 2018, Sydney, Australia, 11/10/18. https://doi.org/10.1109/NSSMIC.2018.8824365
Tao L, Li X, Furenlid LR, Levin CS. Deep Learning Techniques for Gamma Ray Interaction Location Estimation in Monolithic Scintillation Crystal Detectors. In 2018 IEEE Nuclear Science Symposium and Medical Imaging Conference, NSS/MIC 2018 - Proceedings. Institute of Electrical and Electronics Engineers Inc. 2018. 8824365. (2018 IEEE Nuclear Science Symposium and Medical Imaging Conference, NSS/MIC 2018 - Proceedings). https://doi.org/10.1109/NSSMIC.2018.8824365
Tao, Li ; Li, Xin ; Furenlid, Lars R. ; Levin, Craig S. / Deep Learning Techniques for Gamma Ray Interaction Location Estimation in Monolithic Scintillation Crystal Detectors. 2018 IEEE Nuclear Science Symposium and Medical Imaging Conference, NSS/MIC 2018 - Proceedings. Institute of Electrical and Electronics Engineers Inc., 2018. (2018 IEEE Nuclear Science Symposium and Medical Imaging Conference, NSS/MIC 2018 - Proceedings).
@inproceedings{f9ad83557ac447dda1eb756285ed2991,
title = "Deep Learning Techniques for Gamma Ray Interaction Location Estimation in Monolithic Scintillation Crystal Detectors",
abstract = "Monolithic scintillation crystals have been widely studied as detectors for PET systems. In this work, we explore deep learning techniques using the information from mean detector response functions (MDRFs) as a new, potentially faster and more accurate method to estimate gamma ray interaction location in monolithic scintillation crystal detectors. Compared with searching based methods, deep learning techniques do not require recording all the MDRF information once the prediction networks are trained. This could reduce the memory cost by a factor of up to 100. We have designed four different neural networks to estimate/predict gamma ray interaction location given the MDRF data. The networks are trained with calibrated MDRFs at 252×252 locations in a 50.8×50.8 mm2 area, with each MDRF data point containing the information from 20 SiPM channels. Our first network consists of only fully connected (FC) layers, with a final regression layer that directly predicts the values for x and y location. This network is trained to eventually have a loss of 4.8 (L2 loss in units of mm2). The second network is also consisted of only FC layers, but with a final classification layer that classifies the x and y location into 252 classes. This network is trained to have a final prediction accuracy of 15{\%}. The third network is designed as a convolutional neural network (CNN) with a final classification layer. This network is trained to have a final prediction accuracy of 20{\%}. With the last network, we perform {"}sectional training{"} by first coarsely separating the entire crystal into 12×12 sub-areas. Then for each sub-area, we do fine training and classify the x and y location into the correct class. This network is trained to have a final prediction accuracy of 90{\%}. We test the trained networks with a 5 slit image. The RMS prediction errors for the four networks are 2.6 mm (FC regression network), 2.2 mm (FC classification network), 2.1 mm (CNN network) and 2.0 mm ({"}sectional training{"} network). We can see that the CNN network and {"}sectional training{"} network can achieve lower prediction error. In comparison with searching based methods, deep learning based estimation methods do not need to keep record of all the MDRF information, which consists of a total of 1270080 parameters, once the networks are trained. This has reduced the memory cost by a factor of 10 - 100 depending on the network structure.",
keywords = "deep learning, gamma ray interaction location prediction, mean detector response function, Monolithic scintillation detector, neural network",
author = "Li Tao and Xin Li and Furenlid, {Lars R.} and Levin, {Craig S.}",
year = "2018",
month = "11",
doi = "10.1109/NSSMIC.2018.8824365",
language = "English (US)",
series = "2018 IEEE Nuclear Science Symposium and Medical Imaging Conference, NSS/MIC 2018 - Proceedings",
publisher = "Institute of Electrical and Electronics Engineers Inc.",
booktitle = "2018 IEEE Nuclear Science Symposium and Medical Imaging Conference, NSS/MIC 2018 - Proceedings",

}

TY - GEN

T1 - Deep Learning Techniques for Gamma Ray Interaction Location Estimation in Monolithic Scintillation Crystal Detectors

AU - Tao, Li

AU - Li, Xin

AU - Furenlid, Lars R.

AU - Levin, Craig S.

PY - 2018/11

Y1 - 2018/11

N2 - Monolithic scintillation crystals have been widely studied as detectors for PET systems. In this work, we explore deep learning techniques using the information from mean detector response functions (MDRFs) as a new, potentially faster and more accurate method to estimate gamma ray interaction location in monolithic scintillation crystal detectors. Compared with searching based methods, deep learning techniques do not require recording all the MDRF information once the prediction networks are trained. This could reduce the memory cost by a factor of up to 100. We have designed four different neural networks to estimate/predict gamma ray interaction location given the MDRF data. The networks are trained with calibrated MDRFs at 252×252 locations in a 50.8×50.8 mm2 area, with each MDRF data point containing the information from 20 SiPM channels. Our first network consists of only fully connected (FC) layers, with a final regression layer that directly predicts the values for x and y location. This network is trained to eventually have a loss of 4.8 (L2 loss in units of mm2). The second network is also consisted of only FC layers, but with a final classification layer that classifies the x and y location into 252 classes. This network is trained to have a final prediction accuracy of 15%. The third network is designed as a convolutional neural network (CNN) with a final classification layer. This network is trained to have a final prediction accuracy of 20%. With the last network, we perform "sectional training" by first coarsely separating the entire crystal into 12×12 sub-areas. Then for each sub-area, we do fine training and classify the x and y location into the correct class. This network is trained to have a final prediction accuracy of 90%. We test the trained networks with a 5 slit image. The RMS prediction errors for the four networks are 2.6 mm (FC regression network), 2.2 mm (FC classification network), 2.1 mm (CNN network) and 2.0 mm ("sectional training" network). We can see that the CNN network and "sectional training" network can achieve lower prediction error. In comparison with searching based methods, deep learning based estimation methods do not need to keep record of all the MDRF information, which consists of a total of 1270080 parameters, once the networks are trained. This has reduced the memory cost by a factor of 10 - 100 depending on the network structure.

AB - Monolithic scintillation crystals have been widely studied as detectors for PET systems. In this work, we explore deep learning techniques using the information from mean detector response functions (MDRFs) as a new, potentially faster and more accurate method to estimate gamma ray interaction location in monolithic scintillation crystal detectors. Compared with searching based methods, deep learning techniques do not require recording all the MDRF information once the prediction networks are trained. This could reduce the memory cost by a factor of up to 100. We have designed four different neural networks to estimate/predict gamma ray interaction location given the MDRF data. The networks are trained with calibrated MDRFs at 252×252 locations in a 50.8×50.8 mm2 area, with each MDRF data point containing the information from 20 SiPM channels. Our first network consists of only fully connected (FC) layers, with a final regression layer that directly predicts the values for x and y location. This network is trained to eventually have a loss of 4.8 (L2 loss in units of mm2). The second network is also consisted of only FC layers, but with a final classification layer that classifies the x and y location into 252 classes. This network is trained to have a final prediction accuracy of 15%. The third network is designed as a convolutional neural network (CNN) with a final classification layer. This network is trained to have a final prediction accuracy of 20%. With the last network, we perform "sectional training" by first coarsely separating the entire crystal into 12×12 sub-areas. Then for each sub-area, we do fine training and classify the x and y location into the correct class. This network is trained to have a final prediction accuracy of 90%. We test the trained networks with a 5 slit image. The RMS prediction errors for the four networks are 2.6 mm (FC regression network), 2.2 mm (FC classification network), 2.1 mm (CNN network) and 2.0 mm ("sectional training" network). We can see that the CNN network and "sectional training" network can achieve lower prediction error. In comparison with searching based methods, deep learning based estimation methods do not need to keep record of all the MDRF information, which consists of a total of 1270080 parameters, once the networks are trained. This has reduced the memory cost by a factor of 10 - 100 depending on the network structure.

KW - deep learning

KW - gamma ray interaction location prediction

KW - mean detector response function

KW - Monolithic scintillation detector

KW - neural network

UR - http://www.scopus.com/inward/record.url?scp=85073116618&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85073116618&partnerID=8YFLogxK

U2 - 10.1109/NSSMIC.2018.8824365

DO - 10.1109/NSSMIC.2018.8824365

M3 - Conference contribution

AN - SCOPUS:85073116618

T3 - 2018 IEEE Nuclear Science Symposium and Medical Imaging Conference, NSS/MIC 2018 - Proceedings

BT - 2018 IEEE Nuclear Science Symposium and Medical Imaging Conference, NSS/MIC 2018 - Proceedings

PB - Institute of Electrical and Electronics Engineers Inc.

ER -