PET TOF Timing via STFT Waveform Features and Residual CNN
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing PET systems face challenges in achieving precise coincidence time resolution due to errors in detecting gamma photons, including conversion depth, crystal scintillation processes, transmission time, and single photon dispersion, which current timing methods fail to fully utilize the potential information in waveforms.
Innovation Solution
A method utilizing short-time Fourier transform (STFT) and a residual convolutional neural network to process PET detector waveforms, extracting frequency domain features, and predicting Time of Flight (TOF) times, improving coincidence time resolution.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional timing methods (leading edge timing or constant fraction timing) are used to determine photon arrival time from detector waveforms, then the timing process is simple and fast, but the coincidence time resolution is insufficient due to not utilizing potential information in waveforms
Solution Approach 1:
The patent transforms the waveform analysis from time domain to frequency domain using Short-Time Fourier Transform (STFT). This dimensional change allows extraction of frequency domain features that contain additional information about the waveform characteristics, enabling more precise TOF estimation beyond what traditional time domain methods can achieve
Solution Approach 2:
The patent introduces a residual convolutional neural network as an intermediary between the frequency domain features and the TOF estimation. This neural network processes the extracted frequency features and learns optimal mapping relationships, acting as a sophisticated mediator that translates spectral information into accurate timing measurements
2Measurement precision
If deep learning methods are used to extract potential features in waveforms, then the coincidence time resolution can be improved, but the computational complexity and processing time increase
Solution Approach 1:
The patent performs STFT and extracts frequency domain features as preliminary processing steps before TOF estimation. By pre-processing the waveforms into frequency domain representations and extracting relevant features in advance, the actual TOF estimation can be performed more efficiently using the prepared feature sets rather than analyzing raw waveforms in real-time
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
Enhances the temporal resolution of PET systems, enabling more accurate spatial localization and clinical diagnosis by refining TOF information.
Implementation Method 1
the process of crystal scintillation, in which the crystal emits light
Implementation Method 2
converted into electrical signals, inputting the electrical signals into the circuit
Data Source
AI summary
The method for improving the coincidence time resolution of PET system based on STFT, the method obtains waveform data by setting a point source at a specific location, and then obtains a short term frequency domain amplitude information through short time Fourier transform (STFT), after that, carrying out a train set, a validation set, and a test set partitioning, and a residual neural network model composed of residual modules and a fully connected layer is used for training, achieving estimation of the TOF time of the PET system. The present invention utilizes Monte Carlo simulation experiments to perform STFT transform on waveform data to obtain the short term frequency domain amplitude information, then, a residual neural network model is trained by using the divided train set, and the trained model is tested on the validation set to obtain the trained model, finally, the accuracy of TOF time estimation based on STFT model and the improvement effect on the coincidence time resolution of PET system compared to traditional CFD methods are verified through the test set.