Further investigation of 3D dose verification in proton therapy utilizing acoustic signal, wavelet decomposition and machine learning

2021 ◽  
Author(s):  
Songhuan Yao ◽  
Zongsheng Hu ◽  
Qiang Xie ◽  
Yidong Yang ◽  
Hao Peng
Keyword(s):  
Machine Learning ◽  
Proton Therapy ◽  
Acoustic Signal ◽  
Wavelet Decomposition ◽  
Short Term Memory ◽  
Three Dimensions ◽  
Dose Verification ◽  
Case Scenario ◽  
Worst Case ◽  
Dose Distributions

Abstract Online dose verification in proton therapy is a critical task for quality assurance. We further studied the feasibility of using a wavelet-based machine learning framework to accomplishing that goal in three dimensions, built upon our previous work in 1D. The wavelet decomposition was utilized to extract features of acoustic signals and a bidirectional long-short-term memory (Bi-LSTM) recurrent neural network (RNN) was used. The 3D dose distributions of mono-energetic proton beams (multiple beam energies) inside a 3D CT phantom, were generated using Monte-Carlo simulation. The 3D propagation of acoustic signal was modeled using the k-Wave toolbox. Three different beamlets (i.e. acoustic pathways) were tested, one with its own model. The performance was quantitatively evaluated in terms of mean relative error (MRE) of dose distribution and positioning error of Bragg peak (△BP ), for two signal-to-noise ratios (SNRs). Due to the lack of experimental data for the time being, two SNR conditions were modeled (SNR=1 and 5). The model is found to yield good accuracy and noise immunity for all three beamlets. The results exhibit an MRE below 0.6% (without noise) and 1.2% (SNR= 5), and △BP below 1.2 mm (without noise) and 1.3 mm (SNR= 5). For the worst-case scenario (SNR=1), the MRE and △BP are below 2.3% and 1.9 mm, respectively. It is encouraging to find out that our model is able to identify the correlation between acoustic waveforms and dose distributions in 3D heterogeneous tissues, as in the 1D case. The work lays a good foundation for us to advance the study and fully validate the feasibility with experimental results.

A machine learning framework with anatomical prior for online dose verification using positron emitters and PET in proton therapy

2020 ◽  
Vol 65 (18) ◽  
pp. 185003
Author(s):  
Zongsheng Hu ◽  
Guangyao Li ◽  
Xiaoke Zhang ◽  
Kuangkuang Ye ◽  
Jiade Lu ◽  
...  
Keyword(s):  
Machine Learning ◽  
Proton Therapy ◽  
Dose Verification ◽  
Learning Framework ◽  
Positron Emitters

scholarly journals Identification of NLOS and Multi-Path Conditions in UWB Localization Using Machine Learning Methods

2020 ◽  
Vol 10 (11) ◽  
pp. 3980 ◽  
Author(s):  
Cung Lian Sang ◽  
Bastian Steinhagen ◽  
Jonas Dominik Homburg ◽  
Michael Adams ◽  
Marc Hesse ◽  
...  
Keyword(s):  
Machine Learning ◽  
Indoor Localization ◽  
Measurement Data ◽  
Ultra Wideband ◽  
Machine Learning Techniques ◽  
Line Of Sight ◽  
Support Vector ◽  
Case Scenario ◽  
Worst Case ◽  
Worst Case Scenario

In ultra-wideband (UWB)-based wireless ranging or distance measurement, differentiation between line-of-sight (LOS), non-line-of-sight (NLOS), and multi-path (MP) conditions is important for precise indoor localization. This is because the accuracy of the reported measured distance in UWB ranging systems is directly affected by the measurement conditions (LOS, NLOS, or MP). However, the major contributions in the literature only address the binary classification between LOS and NLOS in UWB ranging systems. The MP condition is usually ignored. In fact, the MP condition also has a significant impact on the ranging errors of the UWB compared to the direct LOS measurement results. However, the magnitudes of the error contained in MP conditions are generally lower than completely blocked NLOS scenarios. This paper addresses machine learning techniques for identification of the three mentioned classes (LOS, NLOS, and MP) in the UWB indoor localization system using an experimental dataset. The dataset was collected in different conditions in different scenarios in indoor environments. Using the collected real measurement data, we compared three machine learning (ML) classifiers, i.e., support vector machine (SVM), random forest (RF) based on an ensemble learning method, and multilayer perceptron (MLP) based on a deep artificial neural network, in terms of their performance. The results showed that applying ML methods in UWB ranging systems was effective in the identification of the above-three mentioned classes. Specifically, the overall accuracy reached up to 91.9% in the best-case scenario and 72.9% in the worst-case scenario. Regarding the F1-score, it was 0.92 in the best-case and 0.69 in the worst-case scenario. For reproducible results and further exploration, we provide the publicly accessible experimental research data discussed in this paper at PUB (Publications at Bielefeld University). The evaluations of the three classifiers are conducted using the open-source Python machine learning library scikit-learn.

scholarly journals CBCT-Based Adaptive Assessment Workflow for Intensity Modulated Proton Therapy for Head and Neck Cancer

2021 ◽  
Author(s):  
Mariluz De Ornelas ◽  
Yihang Xu ◽  
Kyle Padgett ◽  
Ryder M. Schmidt ◽  
Michael Butkus ◽  
...  
Keyword(s):  
Computed Tomography ◽  
Head And Neck ◽  
Proton Therapy ◽  
Case Scenario ◽  
Worst Case ◽  
Intensity Modulated Proton Therapy ◽  
Intensity Modulated ◽  
Worst Case Scenario ◽  
Patient Setup ◽  
Delivered Dose

Abstract Purpose Anatomical changes and patient setup uncertainties during intensity modulated proton therapy (IMPT) of head and neck (HN) cancers demand frequent evaluation of delivered dose. This work investigated a cone-beam computed tomography (CBCT) and deformable image registration based therapy workflow to demonstrate the feasibility of proton dose calculation on synthetic computed tomography (sCT) for adaptive IMPT treatment of HN cancer. Materials and Methods Twenty-one patients with HN cancer were enrolled in this study, a retrospective institutional review board protocol. They had previously been treated with volumetric modulated arc therapy and had daily iterative CBCT. For each patient, robust optimization (RO) IMPT plans were generated using ±3 mm patient setup and ±3% proton range uncertainties. The sCTs were created and the weekly delivered dose was recalculated using an adaptive dose accumulation workflow in which the planning computed tomography (CT) was deformably registered to CBCTs and Hounsfield units transferred from the planning CT. Accumulated doses from ±3 mm/±3% RO-IMPT plans were evaluated using clinical dose-volume constraints for targets (clinical target volume, or CTV) and organs at risk. Results Evaluation of weekly recalculated dose on sCTs showed that most of the patient plans maintained target dose coverage. The primary CTV remained covered by the V95 > 95% (95% of the volume receiving more than 95% of the prescription dose) worst-case scenario for 84.5% of the weekly fractions. The oral cavity accumulated mean dose remained lower than the worst-case scenario for all patients. Parotid accumulated mean dose remained within the uncertainty bands for 18 of the 21 patients, and all were kept lower than RO-IMPT worst-case scenario for 88.7% and 84.5% for left and right parotids, respectively. Conclusion This study demonstrated that RO-IMPT plans account for most setup and anatomical uncertainties, except for large weight-loss changes that need to be tracked throughout the treatment course. We showed that sCTs could be a powerful decision tool for adaptation of these cases in order to reduce workload when using repeat CTs.

Linear Growth Rate Estimation From Dynamics and Statistics of Acoustic Signal Envelope in Turbulent Combustors

2016 ◽  
Vol 139 (4) ◽  
Author(s):  
Nicolas Noiray
Keyword(s):  
Gas Turbine ◽  
Acoustic Signal ◽  
Growth Rates ◽  
Linear Growth ◽  
Dynamic Pressure ◽  
Linear Growth Rate ◽  
Case Scenario ◽  
Major Step ◽  
Worst Case ◽  
Signal Envelope

Considerable research and development efforts are required to meet the targets of future gas turbine technologies in terms of performance, emissions, and operational flexibility. One of the recurring problems is the constructive coupling between flames and combustor's acoustics. These thermoacoustic interactions can cause high-amplitude dynamic pressure limit cycles, which reduce the lifetime of the hot gas path parts or in the worst-case scenario destroy these mechanical components as a result of a sudden catastrophic event. It is shown in this paper that the dynamics and the statistics of the acoustic signal envelope can be used to identify the linear growth rates hidden behind the observed pulsations, and the results are validated against numerical simulations. This is a major step forward and it will contribute to the development of future gas turbine combustors, because the knowledge of these linear growth rates is essential to develop robust active and passive systems to control these combustion instabilities.

Machine Learning Prediction and Reliability Analysis Applied to Subsea Spool and Jumper Design

2021 ◽  
Author(s):  
Mengdi Song ◽  
Massyl Gheroufella ◽  
Paul Chartier
Keyword(s):  
Machine Learning ◽  
Reliability Analysis ◽  
Design Process ◽  
Predictive Modelling ◽  
Machine Learning Algorithms ◽  
Case Scenario ◽  
Worst Case ◽  
Deterministic Solution ◽  
Number Of Bends ◽  
Subsea Pipelines

Abstract In subsea pipelines projects, the design of rigid spool and jumper can be a challenging and time-consuming task. The selected spool layout for connecting the pipelines to the subsea structures, including the number of bends and leg lengths, must offer the flexibility to accommodate the pipeline thermal expansion, the pipe-lay target box and misalignments associated with the post-lay survey metrology and spool fabrication. The analysis results are considerably affected by many uncertainties involved. Consequently, a very large amount of calculations is required to assess the full combination of uncertainties and to capture the worst-case scenario. Rather than applying the deterministic solution, this paper uses machine learning prediction to significantly improve the efficiency of the design process. In addition, thanks to the fast predictive model using machine learning algorithms, the uncertainty quantification and propagation analysis using probabilistic statistical method becomes feasible in terms of CPU time and can be incorporated into the design process to evaluate the reliability of the outputs. The latter allows us to perform a systematic probabilistic design by considering a certain level of acceptance on the probability of failure, for example as per DNVGL design code. The machine learning predictive modelling and the reliability analysis based upon the probability distribution of the uncertainties are introduced and explained in this paper. Some project examples are shown to highlight the method’s comprehensive nature and efficient characteristics.

scholarly journals Identification of NLOS and Multi-Path Conditions in UWB Localization Using Machine Learning Methods

2020 ◽  
Author(s):  
Cung Lian Sang ◽  
Bastian Steinhagen ◽  
Jonas Dominik Homburg ◽  
Michael Adams ◽  
Marc Hesse ◽  
...  
Keyword(s):  
Machine Learning ◽  
Indoor Localization ◽  
Measurement Data ◽  
Ultra Wideband ◽  
Line Of Sight ◽  
Support Vector ◽  
Case Scenario ◽  
Worst Case ◽  
Data Set ◽  
Worst Case Scenario

In Ultra-wideband (UWB)-based wireless ranging or distance measurement, differentiation between line-of-sight~(LOS), non-line-of-sight~(NLOS), and multi-path (MP) conditions are important for precise indoor localization. This is because the accuracy of the reported measured distance in UWB ranging systems is directly affected by the measurement conditions (LOS, NLOS or MP). However, the major contributions in literature only address the binary classification between LOS and NLOS in UWB ranging systems. The MP condition is usually ignored. In fact, the MP condition also has a significant impact on the ranging errors of the UWB compared to the direct LOS measurement results. Though, the magnitudes of the error contained in MP conditions are generally lower than completely blocked NLOS scenarios. This paper addresses machine learning techniques for identification of the mentioned three classes (LOS, NLOS, and MP) in the UWB indoor localization system using an experimental data-set. The data-set was collected in different conditions at different scenarios in indoor environments. Using the collected real measurement data, we compare three machine learning (ML) classifiers, i.e., support vector machine (SVM), random forest (RF) based on an ensemble learning method, and multilayer perceptron (MLP) based on a deep artificial neural network, in terms of their performance. The results show that applying ML methods in UWB ranging systems are effective in identification of the above-mentioned three classes. In specific, the overall accuracy reaches up to 91.9% in the best-case scenario and 72.9% in the worst-case scenario. Regarding the F1-score, it is 0.92 in the best-case and 0.69 in the worst-case scenario. For reproducible results and further exploration, we (will) provide the publicly accessible experimental research data discussed in this paper at PUB - Publications at Bielefeld University. The evaluations of the three classifiers are conducted using the open-source python machine learning library scikit-learn.

scholarly journals A New Constructive Heuristic Driven by Machine Learning for the Traveling Salesman Problem

Algorithms ◽  
2021 ◽  
Vol 14 (9) ◽  
pp. 267
Author(s):  
Umberto Junior Mele ◽  
Luca Maria Gambardella ◽  
Roberto Montemanni
Keyword(s):  
Machine Learning ◽  
Traveling Salesman Problem ◽  
Scale Up ◽  
Search Space ◽  
Optimization Techniques ◽  
Traveling Salesman ◽  
Case Scenario ◽  
Worst Case ◽  
Constructive Heuristic ◽  
The Traveling Salesman Problem

Recent systems applying Machine Learning (ML) to solve the Traveling Salesman Problem (TSP) exhibit issues when they try to scale up to real case scenarios with several hundred vertices. The use of Candidate Lists (CLs) has been brought up to cope with the issues. A CL is defined as a subset of all the edges linked to a given vertex such that it contains mainly edges that are believed to be found in the optimal tour. The initialization procedure that identifies a CL for each vertex in the TSP aids the solver by restricting the search space during solution creation. It results in a reduction of the computational burden as well, which is highly recommended when solving large TSPs. So far, ML was engaged to create CLs and values on the elements of these CLs by expressing ML preferences at solution insertion. Although promising, these systems do not restrict what the ML learns and does to create solutions, bringing with them some generalization issues. Therefore, motivated by exploratory and statistical studies of the CL behavior in multiple TSP solutions, in this work, we rethink the usage of ML by purposely employing this system just on a task that avoids well-known ML weaknesses, such as training in presence of frequent outliers and the detection of under-represented events. The task is to confirm inclusion in a solution just for edges that are most likely optimal. The CLs of the edge considered for inclusion are employed as an input of the neural network, and the ML is in charge of distinguishing when such edge is in the optimal solution from when it is not. The proposed approach enables a reasonable generalization and unveils an efficient balance between ML and optimization techniques. Our ML-Constructive heuristic is trained on small instances. Then, it is able to produce solutions—without losing quality—for large problems as well. We compare our method against classic constructive heuristics, showing that the new approach performs well for TSPLIB instances up to 1748 cities. Although ML-Constructive exhibits an expensive constant computation time due to training, we proved that the computational complexity in the worst-case scenario—for the solution construction after training—is O(n2logn2), n being the number of vertices in the TSP instance.

Sign in / Sign up

Export Citation Format

Share Document