LZGD: Massively parallel GPU based compression mechanism

TECHNOLOGY NUMBER: 2023-293
Technology No. 2023-293

OVERVIEW  

GPU-based image compression and buffering technology for whole slide digital pathology.

  • Replaces slow, rigid processes with real-time tile buffering and single-action GPU parallel compression for lower latency.
  • Digital pathology labs, real-time clinical consultations, AI workflows, geospatial imaging, and remote sensing.

BACKGROUND 

Whole slide imaging is transforming pathology and medical diagnostics, yet handling massive digital slide files presents significant challenges. Traditional imaging systems rely on slow, sequential processing and heuristic assumptions that introduce high latency, slow file retrieval, and delays in rendering. Furthermore, current solutions often depend on rigid proprietary formats that complicate software integration, restrict interoperability across vendors, and hinder data sharing across networks. These inefficiencies limit the performance of high-throughput clinical workflows, real-time remote consultations, and AI-driven image analysis. Beyond healthcare, fields like geospatial imaging and remote sensing face similar data reduction and retrieval bottlenecks. Consequently, there is a critical need for an open, scalable, and vendor-neutral technology capable of rapidly compressing, buffering, and serving massive high-resolution images with minimal latency and superior responsiveness.

INNOVATION

Iris and LZGD combine to deliver a next-generation GPU-based image compression and buffering platform. Iris employs rapid tile buffering that actively responds to user input, dynamically loading required image data to provide instant responses, higher fidelity scaling, and simple error recovery. LZGD utilizes the parallel power of modern GPUs to detect repeated patterns across entire images in a single action, drastically accelerating compression speed and data reduction. Complementing this performance, Iris introduces an open, modern binary file format offering state-of-the-art compression, instant tile access, built-in validation, and easy integration across software languages and platforms. This vendor-agnostic technology transforms whole slide imaging, enabling high-throughput digital pathology labs, real-time clinical collaborations, and seamless integration into AI medical workflows, with broader utility in geospatial imaging and remote sensing.


ADDITIONAL INFORMATION

INTELLECTUAL PROPERTY

Pending


RELATED TECHNOLOGY


  • expand_more mode_edit Inventor (1)
    Ryan Landvater
  • expand_more cloud_download Supporting documents (1)
    Product brochure
    LZGD: Massively parallel GPU based compression mechanism.pdf
Questions about this technology?