Dynamic Image Temporal Interpolation Using Deformation Encoding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing dynamic imaging modalities face challenges in achieving high temporal resolution without compromising spatial resolution, signal-to-noise ratio, or scan time, which impacts clinical utility in applications such as cardiac MRI.
Innovation Solution
A deformation encoding neural network is used to derive parameters characterizing image dynamics and generate interpolated frames, increasing the frame rate of dynamic images without altering spatial resolution or scan time, by integrating with existing imaging protocols as a post-processing tool.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If image acceleration is used to achieve high spatial and temporal resolution, then temporal resolution is improved, but image quality and signal-to-noise ratio are compromised
Solution Approach 1:
The system performs preliminary action by using a neural network to predict and generate interpolated image frames between actual acquired frames. The neural network is trained to anticipate the content of intermediate frames based on temporal patterns, allowing the system to synthesize high-quality intermediate frames without requiring additional actual acquisitions that would compromise SNR.
Solution Approach 2:
The system creates copies of image frames through neural network interpolation. Instead of acquiring new frames that would increase scan time and reduce SNR, the system generates synthetic copies of intermediate frames by copying and transforming features from adjacent actual frames through the neural network's deformation encoding.
2Measurement precision
If frame rate is increased to improve temporal resolution, then temporal resolution is improved, but spatial resolution and scan time are compromised
Solution Approach 1:
The neural network performs preliminary analysis of temporal patterns and deformation fields from actual frames, allowing the system to predict intermediate frames without requiring additional actual acquisitions. This preliminary computation enables high frame rate synthesis while preserving spatial resolution of the original acquired frames.
Solution Approach 2:
The neural network acts as an intermediary between actual image frames and the desired high-frame-rate output. It processes the deformation encoding and temporal patterns from actual frames to generate interpolated frames, serving as a mediator that bridges the gap between low frame rate acquisition and high frame rate output without sacrificing spatial quality.
3Measurement precision
If scan time is increased to achieve higher temporal resolution, then temporal resolution is improved, but productivity and efficiency are reduced
Solution Approach 1:
The system performs all necessary temporal interpolation computations after the actual scan is complete, using the neural network to generate additional frames from the acquired data. This eliminates the need to extend scan time during acquisition, maintaining productivity while achieving high temporal resolution through post-processing.
Solution Approach 2:
The neural network continuously processes the deformation encoding and generates interpolated frames in a continuous computational workflow after acquisition. This continuous processing allows the system to maintain high frame rate output without interrupting or extending the actual scan time, preserving scan efficiency.
Data Source
AI summary
A system for increasing a frame rate of a dynamic image includes an input for receiving a set of consecutive image frames of a first dynamic image having a first plurality of image frames and a first frame rate. The system further includes a deformation encoding neural network coupled to the input and configured to derive at least one parameter characterizing dynamics of the set of consecutive image frames and to generate an interpolated image frame based on the at least one parameter, and a post-processing module coupled to the deformation encoding neural network and configured to receive one or more interpolated image frames from the deformation encoding neural network, to create a second plurality of image frames comprising the one or more interpolated image frames and the first plurality of image frames, and to generate a second dynamic image using the second plurality of image frames, the second dynamic image having a second frame rate higher than the first frame rate.