Deinterlacing via Equal-Intensity Path Recognition

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing deinterlacing methods often result in image artifacts like 'jaggies' due to simple interpolation and require large search areas for edge detection, degrading the ability to resolve fine image details.

Innovation Solution

The method employs equal-intensity lines to interpolate missing sample values, recognizing and partially straightening equal-intensity paths to determine intensity values for missing samples, reducing the need for large search areas and improving image detail resolution.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If simple interpolation is used to obtain missing sample values, then the deinterlacing process is simple and fast, but image artifacts like 'jaggies' occur and image quality deteriorates

Engineering Contradiction:

Improvedeinterlacing processing speedVSAvoidimage quality

Core Design Contradiction:

ProductivityVSManufacturing precision

Solution Approach 1:

The patent applies different processing strategies to different regions of the image based on local characteristics. Equal-intensity paths are identified and treated differently from regular regions, with path recognition and straightening applied selectively to preserve edge quality while maintaining efficiency in uniform areas.

Inventive Principle:

Principle #3Local quality

Solution Approach 2:

The patent transitions from traditional spatial interpolation methods to a method based on equal-intensity paths in the intensity domain. By tracking paths of constant intensity values and straightening them, the method operates in a transformed space that preserves edge information better than conventional spatial interpolation.

Inventive Principle:

Principle #17Another dimension (Dimensionality change)

2Measurement precision

If large search areas are used for edge detection, then edge detection accuracy improves, but the ability to resolve fine image details deteriorates

Engineering Contradiction:

Improveedge detection accuracyVSAvoidfine image detail resolution

Core Design Contradiction:

Measurement precisionVSManufacturing precision

Solution Approach 1:

The patent segments the image processing into distinct stages: equal-intensity trace generation, path recognition, path straightening, and intensity value determination. This segmentation allows each stage to operate with optimized parameters, avoiding the need for large search areas while maintaining edge detection accuracy.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The patent performs preliminary generation of equal-intensity traces from present samples before proceeding to path recognition and straightening. This preliminary action prepares the data in a form that makes subsequent edge detection more accurate without requiring large search areas, as the equal-intensity traces already encode edge information.

Inventive Principle:

Principle #10Preliminary action

Data Source

PatentUS9549147B2System and method for creating a video frame from a single video field

Publication Date: 2017.01.17 NVIDIA CORP

AI summary

A system and method of producing a frame of a video image from an interlaced field. In one embodiment, the method includes: (1) creating an equal-intensity trace from present samples in the field, (2) recognizing an equal-intensity path in the equal-intensity trace, (3) at least partially straightening the equal-intensity path and (4) using the equal-intensity path to determine an intensity value for a missing sample in the frame.