Distributed image processing method, apparatus, device, storage medium, and program product

CN122244253APending Publication Date: 2026-06-19INDUSTRIAL AND COMMERCIAL BANK OF CHINA

Patent Information

Authority / Receiving Office: CN · China
Patent Type: Applications(China)
Current Assignee / Owner: INDUSTRIAL AND COMMERCIAL BANK OF CHINA
Filing Date: 2026-01-21
Publication Date: 2026-06-19

AI Technical Summary

Technical Problem

In the distributed batch processing scenarios of fintech, existing technologies face a conflict between data privacy and processing efficiency in the visualization of graph data, and an imbalance between rendering performance and information fidelity, resulting in low image processing efficiency and difficulty in visualizing massive graph data.

Method used

By employing federated graph dimensionality reduction technology, fused feature vectors are generated locally on distributed data nodes. Global low-dimensional coordinates are generated through gradient information aggregation. Combined with multi-level detail models and intelligent sharding strategies, efficient collaborative processing is achieved, avoiding the transmission of raw data across nodes and meeting privacy protection requirements.

Benefits of technology

Significantly reduces data transmission volume, shortens preprocessing time, improves computational efficiency and privacy security, enhances rendering performance and interactive response speed, and strengthens the privacy security and computational efficiency of image data processing.

✦ Generated by Eureka AI based on patent content.

Smart Images

Figure CN122244253A_ABST

Patent Text Reader

Abstract

This application provides a distributed image processing method, apparatus, device, storage medium, and program product, relating to the fields of big data and fintech. The method includes: acquiring image data to be processed; performing local feature calculations on the image data in distributed data nodes to generate a fused feature vector corresponding to the image data; obtaining gradient information of the local distance matrix corresponding to the fused feature vector based on the fused feature vector, and aggregating it in a central coordination node to generate global low-dimensional coordinates corresponding to the image data; and in a rendering cluster node, performing multi-level detail model generation based on the global low-dimensional coordinates, and combining user interaction commands to obtain the rendering result corresponding to the image data. The distributed data nodes, central coordination node, and rendering cluster nodes are interconnected via a network using a point-to-point communication protocol. This method simultaneously improves the privacy and security as well as the computational efficiency of cross-node image data processing.

Need to check novelty before this filing date? Find Prior Art