Real-time hardware-assisted gpu tuning using machine learning
US-2019213775-A1 · Jul 11, 2019 · US
US10497477B2 · US · B2
| Field | Value |
|---|---|
| Publication number | US-10497477-B2 |
| Application number | US-201515507444-A |
| Country | US |
| Kind code | B2 |
| Filing date | Aug 28, 2015 |
| Priority date | Aug 29, 2014 |
| Publication date | Dec 3, 2019 |
| Grant date | Dec 3, 2019 |
A practical reading order for non-experts. Skip the full description unless you need deep technical detail.
What the patent document calls the invention.
A short plain-language summary of the technical disclosure.
Who owns or filed the patent and who is credited as inventor.
Filing, priority, publication, and grant dates set the timeline.
The legal scope of protection — read this for what is actually claimed.
Technology tags used to group this patent with similar filings.
Prior art links and similar publications in this corpus.
Official abstract text for this publication.
The present invention relates to a method for high-speed parallel processing for an ultrasonic signal, the method used for generation of an ultrasonic image by a smart device, which is provided with a mobile graphic processing unit (GPU), by receiving an input of an ultrasonic signal. The method comprises the steps of: receiving an input of an ultrasonic signal beam-formed by means of a first rendering cycle, removing a DC component from the ultrasonic signal, and then separating an in-phase component and a quadrature component from the ultrasonic signal, from which the DC component has been removed, and separately outputting same; a smart device performing quadrature demodulation and envelope detection processing for the ultrasonic signal, having the in-phase component and the quadrature component, by means of a second rendering cycle; and the smart device performing scan conversion for the ultrasonic signal, which has been obtained as the result of the second rendering cycle, by means of a fifth rendering cycle, wherein the rendering cycles are formed as a graphics pipeline structure comprising a vertex shader procedure, a rasterizer procedure, and a fragment shader procedure. A method for high-speed parallel processing for an ultrasonic signal by using a smart device, according to the present invention, enables high-speed parallel processing for an ultrasonic signal by means of a mobile GPU inside a smart device even in a mobile-based environment instead of a PC-based environment, thereby enabling the providing of an image having a frame rate that is useful for medical diagnosis.
Opening claim text (preview).
The invention claimed is: 1. A method for high-speed parallel processing for an ultrasonic signal by using a smart device, which is used for a smart device having a mobile GPU (Graphic Processing Unit) to receive an ultrasonic signal and generate an ultrasonic image, the method comprising: by the smart device, receiving a beam-formed ultrasonic signal through a first render cycle, removing a DC component from the ultrasonic signal, and then dividing and outputting an in-phase component and a quadrature component from the ultrasonic signal free from the DC component; by the smart device, performing quadrature demodulation processing and envelope detection processing through a second render cycle to the ultrasonic signal having the in-phase component and the quadrature component; and by the smart device, performing scan conversion through a fifth render cycle to the ultrasonic signal obtained as a result of the second render cycle, wherein one or more of the render cycles has a graphics pipeline structure including a vertex shader stage, a rasterizer stage and a fragment shader stage, wherein the mobile GPU controls a part of operations allocated to the fragment shader stage to be performed in advance in the vertex shader stage. 2. The method for high-speed parallel processing for an ultrasonic signal by using a smart device according to claim 1 , wherein in the vertex shader stage, the mobile GPU receives a plurality of vertexes, allocates a spatial region by using the received vertexes, and then calculates a spatial coordinate for the allocated spatial region to generate a calculation result in the form of a varying parameter. 3. The method for high-speed parallel processing for an ultrasonic signal by using a smart device according to claim 2 , wherein in the rasterizer stage, the mobile GPU searches an on-screen coordinate value corresponding to the varying parameter output in the vertex shader stage, and generates the searched coordinate value in the form of a varying parameter. 4. The method for high-speed parallel processing for an ultrasonic signal by using a smart device according to claim 3 , wherein in the vertex shader stage, the mobile GPU further calculates a coordinate value located on the periphery of the on-screen coordinate value generated in the rasterizer stage, and generates a calculation result in the form of a varying parameter. 5. The method for high-speed parallel processing for an ultrasonic signal by using a smart device according to claim 4 , wherein in the fragment shader stage, the mobile GPU calculates a color for the on-screen coordinate value generated in the rasterizer stage and generates a calculation result. 6. The method for high-speed parallel processing for an ultrasonic signal by using a smart device according to claim 5 , further comprising: by the mobile GPU, storing the color calculation result for the on-screen coordinate value generated in the fragment shader stage in a frame buffer. 7. The method for high-speed parallel processing for an ultrasonic signal by using a smart device according to claim 2 , wherein 8 to 16 varying parameters are output depending on a specification of the mobile GPU. 8. The method for high-speed parallel processing for an ultrasonic signal by using a smart device according to claim 1 , wherein when an image generated in a previous render cycle is stored in a frame buffer, the mobile GPU transfers the stored image to a texture corresponding to a memory of the mobile GPU by means of a RTT (Render To Texture) technique, and transfers the transferred image to a fragment shader stage in a graphics pipeline of a next render cycle. 9. The method for high-speed parallel processing for an ultrasonic signal by using a smart device according to claim 1 , wherein the mobile GPU performs parallel processing to an ultrasonic signal by using a fixed function pipeline structure. 10. The method for high-speed parallel processing for an ultrasonic signal by using a smart device according to claim 1 , wherein the method is implemented under OpenGL ES environment. 11. The method for high-speed parallel processing for an ultrasonic signal by using a smart device according to claim 1 , wherein in the second render cycle, the smart device further performs decimation processing to the ultrasonic signal to which the quadrature demodulation processing is performed. 12. The method for high-speed parallel processing for an ultrasonic signal by using a smart device according to claim 1 , wherein in the second render cycle, the smart device further performs log compression processing to the ultrasonic signal to which the envelope detection processing is performed. 13. The method for high-speed parallel processing for an ultrasonic signal by using a smart device according to claim 12 , wherein in the second render cycle, the smart device further performs gain control processing to control an overall gain of an image with respect to the ultrasonic signal to which the log compression processing is performed. 14. The method for high-speed parallel processing for an ultrasonic signal by using a smart device according to claim 1 , wherein after the second render cycle is performed, the smart device further performs removing a blackhole for the ultrasonic signal by receiving a threshold value which is regarded as a blackhole through a third render cycle and comparing sizes of the threshold value and the ultrasonic signal received in the second render cycle with each other. 15. The method for high-speed parallel processing for an ultrasonic signal by using a smart device according to claim 1 , wherein after the second render cycle is performed, the smart device further performs edge enhancing through a fourth render cycle to the ultrasonic signal received in the second render cycle. 16. A non-transitory computer-readable recording medium, on which a program for executing the method defined in claim 1 with a computer is recorded. 17. A method for high-speed parallel processing for an ultrasonic signal by using a smart device, which is used for a smart device having a mobile GPU (Graphic Processing Unit) to receive an ultrasonic signal and generate an ultrasonic image, the method comprising: by the smart device, receiving a beam-formed ultrasonic signal through a first render cycle, removing a DC component from the ultrasonic signal, and then dividing and outputting an in-phase component and a quadrature component from the ultrasonic signal free from the DC component; by the smart device, performing quadrature demodulation processing and envelope detection processing through a second render cycle to the ultrasonic signal having the in-phase component and the quadrature component; and by the smart device, performing scan conversion through a fifth render cycle to the ultrasonic signal obtained as a result of the second render cycle, wherein one or more of the render cycles has a graphics pipeline structure including a vertex shader stage, a rasterizer stage and a fragment shader stage, and wherein after the second render cycle is performed, the smart device further performs removing a blackhole for the ultrasonic signal by receiving a threshold value which is regarded as a blackhole through a third render cycle and comparing sizes of the threshold value and the ultrasonic signal received in the second render cycle with each other. 18. The method for high-speed parallel processing for an ultrasonic signal by using a smart device according to claim 17 , wherein after the second render cycle is performed, the smart device further per
using special techniques for image reconstruction, e.g. FFT, geometrical transformations, spatial deconvolution, time deconvolution · CPC title
Enhancement of operational speed, e.g. by using several microcontrol devices operating in parallel · CPC title
involving processing of raw data to produce diagnostic data, e.g. for generating an image · CPC title
using quadrature demodulation · CPC title
for elimination of side lobes or of grating lobes; for increasing resolving power · CPC title
Related publications grouped by family.
Answers are generated from the same data shown on this page.