Distributed OCT Processing with GPU Offload

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Current OCT imaging systems face challenges in acquiring, processing, and displaying high-speed data at real-time frame rates, primarily due to the CPU's burden of performing all Fast Fourier Transform (FFT) calculations, which limits their ability to handle the high bit rates generated by high-speed imaging.

Innovation Solution

A distributed computing system is employed, utilizing multiple networked computers or processors to distribute OCT steps such as data acquisition, pre-processing, Fourier transformation, post-processing, and display, with the option to offload CPU-intensive tasks like FFT calculations to a Graphics Processing Unit (GPU) or Field Programmable Gate Array (FPGA), enabling efficient processing and real-time image generation.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If the CPU performs all FFT calculations in high-speed OCT imaging systems, then the system can maintain a simple single-processor architecture, but the processing speed and real-time frame rate capability are limited due to the CPU's computational burden

Engineering Contradiction:

Improveprocessing speedVSAvoidsystem architecture complexity

Core Design Contradiction:

ProductivityVSDevice complexity

Solution Approach 1:

The patent divides the FFT processing workload into multiple segments handled by different processors. Specifically, it uses a multi-processor system where each processor handles a portion of the spectral data, performing parallel FFT calculations to achieve higher overall processing throughput than a single CPU could provide

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The patent replaces the general-purpose CPU-based processing system with specialized hardware accelerators (GPUs or FPGAs) that are optimized for parallel computational tasks like FFT. This substitution provides significant speedup while managing complexity through dedicated hardware design

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

2Loss of time

If a distributed computing system with multiple processors is used to handle FFT calculations, then real-time processing capability is improved, but the system architecture becomes more complex requiring coordination between multiple processors

Engineering Contradiction:

Improveprocessing timeVSAvoidprocessor coordination complexity

Core Design Contradiction:

Loss of timeVSDevice complexity

Solution Approach 1:

The patent segments the input spectral data into distinct portions assigned to different processors. Each processor independently performs FFT on its assigned segment, and the results are subsequently combined. This segmentation enables parallel processing that reduces overall processing time while maintaining manageable complexity through clear division of labor

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The patent performs preliminary data preparation and segmentation before distributing to multiple processors. By pre-organizing the spectral data into processable segments and establishing the distribution scheme in advance, the system minimizes coordination overhead during actual processing, thus reducing processing time without proportionally increasing complexity

Inventive Principle:

Principle #10Preliminary action

Data Source

PatentEP2201325B1Real time spectral-domain optical coherence tomography (sd-oct) with distributed acquisition and processing

Publication Date: 2019.04.17 PHILIPS IMAGE GUIDED THERAPY CORP

Figure 1
Figure 2A
Figure 2B

AI summary

The present invention relates to distributed computing system to acquire, process, and display OCT images at real-time frame rates.