Mobile GPU Ultrasonic Signal Processing via Graphics Pipeline

The patent segments the ultrasonic signal processing into multiple render cycles, with each cycle handling specific processing stages (beam-forming, DC removal, quadrature demodulation, envelope detection, scan conversion). This segmentation allows the GPU to process different aspects of the signal in parallel, significantly improving processing speed while maintaining manageable complexity through structured organization of processing tasks.

The patent replaces the traditional CPU-based sequential processing mechanism with a GPU-based parallel processing mechanism. By utilizing the graphics pipeline structure and shader programs (vertex shader, fragment shader), the system substitutes the mechanical sequential execution model with a parallel computation model, achieving the necessary processing speed for medical diagnosis while maintaining portability.

The patent utilizes the graphics pipeline structure, which is a universal interface available across different GPU architectures and mobile platforms. By implementing ultrasonic signal processing through standard graphics shader programs (vertex shader, fragment shader) rather than architecture-specific APIs like CUDA or OpenCL, the system achieves high processing efficiency while maintaining broad compatibility across mobile devices with different GPU vendors.

The patent adapts the processing algorithm to the mobile GPU environment by changing the implementation parameters from PC-based APIs to mobile graphics pipeline parameters. This includes using render cycles, shader programs, and graphics memory structures that are native to mobile GPU architectures, thereby achieving both high productivity and environment compatibility.

The patent segments the computational load by distributing calculations across different shader stages. The vertex shader handles spatial coordinate calculations and allocates spatial regions, while the fragment shader focuses on pixel-level processing such as envelope detection and scan conversion. This segmentation maintains processing accuracy by ensuring each stage performs its designated calculations precisely while improving overall speed through parallel execution across stages.

The patent performs preliminary calculations in the vertex shader stage, such as allocating spatial regions and calculating spatial coordinates before the fragment shader stage processes the actual ultrasonic signal data. This preliminary action reduces the computational burden on the fragment shader, thereby improving processing speed while maintaining accuracy as the fragment shader receives pre-processed data that requires fewer calculations.