Parallel Pixel Extraction for Image Processing Throughput

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing technologies for deep learning, particularly in the context of image processing, face challenges in efficiently performing pixel extraction and weight gradient calculation, which hampers overall computing throughput and user experience.

Innovation Solution

The proposed method utilizes dedicated processing units to perform parallel pixel extraction on each row of pixels in an input image, followed by stitching the row data to obtain an output image, thereby enhancing the efficiency of weight gradient calculation and improving overall computing throughput.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If parallel pixel extraction using multiple dedicated processing units is implemented, then computing throughput is improved, but device complexity increases

Engineering Contradiction:

Improvecomputing throughputVSAvoiddevice complexity

Core Design Contradiction:

ProductivityVSDevice complexity

Solution Approach 1:

The image processing task is segmented into row-wise pixel extraction operations that can be independently executed by multiple dedicated processing units in parallel, thereby increasing computing throughput while maintaining manageable device complexity through functional decomposition

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The patent introduces a new dimension of parallelism by processing image rows simultaneously across multiple dedicated processing units rather than sequential pixel processing, achieving throughput improvement through spatial parallelization

Inventive Principle:

Principle #17Another dimension (Dimensionality change)

2Productivity

If efficient weight gradient calculation is implemented, then productivity is improved, but ease of operation deteriorates

Engineering Contradiction:

Improveweight gradient calculation efficiencyVSAvoidease of operation

Core Design Contradiction:

ProductivityVSEase of operation

Solution Approach 1:

The patent introduces dedicated processing units as intermediary components that handle the complex weight gradient calculation operations, shielding the user from complexity while achieving efficient computation through specialized hardware acceleration

Inventive Principle:

Principle #24Intermediary (Mediator)

Data Source

PatentUS12288373B2Method of processing image, electronic device, storage medium, and program product

Publication Date: 2025.04.29 KUNLUNXIN TECHNOLOGY (BEIJING) CO LTD

AI summary

The present disclosure provides a method of processing an image, an electronic device, and a storage medium, which may be used in a field of artificial intelligence, especially in a field of image processing, etc. The method includes: acquiring an input image containing a plurality of rows of pixels; performing, by using a plurality of dedicated processing units, a pixel extraction in parallel on each row of pixels of the plurality of rows of pixels in the input image, so as to obtain row data for each row of pixels; and stitching the row data for each row of pixels, so as to obtain an output image.