Adaptive Geometric Contouring for Vertex-Minimized Concave Hulls

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Conventional concave hull algorithms are approximate and sacrifice region integrity for spatial and temporal efficiency, failing to accurately represent real-world shapes in applications like autonomous driving and biomedical imaging.

Innovation Solution

The adaptive geometric contouring (AGC) algorithm generates vertex-minimized concave hulls with pixel-perfect resolution, using a deterministic approach to identify contour edges and vertices, ensuring maximal data integrity and flexibility in speed-efficiency tradeoffs.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If conventional concave hull algorithms are used, then spatial and temporal efficiency is improved, but region integrity and shape accuracy deteriorate

Engineering Contradiction:
Improvespatial and temporal efficiencyVSAvoidshape accuracy
Core Design Contradiction:
ProductivityVSManufacturing precision

Solution Approach 1:

The patent changes the fundamental parameters of contour generation by using a deterministic approach that evaluates multiple candidate edges at each vertex and selects the optimal edge based on geometric criteria. This transforms the approximate nature of conventional algorithms into a precise method that maintains region integrity while still achieving computational efficiency through systematic edge evaluation and selection.

Inventive Principle:
Principle #35Parameter changes

Solution Approach 2:

The patent introduces dynamic edge selection where the contour edges are not fixed in advance but are determined adaptively during the algorithm execution. At each vertex, the algorithm dynamically evaluates multiple candidate edges and selects the best one based on geometric properties, allowing the contour to adapt to the actual shape features while maintaining efficiency.

Inventive Principle:
Principle #15Dynamics

2Loss of time

If approximate concave hull algorithms are used, then computational speed is improved, but data integrity deteriorates

Engineering Contradiction:
Improvecomputational speedVSAvoiddata integrity
Core Design Contradiction:
Loss of timeVSLoss of information

Solution Approach 1:

The patent replaces the approximate mechanical sampling approach with a deterministic geometric system. Instead of using rough approximations that sacrifice data integrity, the system systematically evaluates geometric properties of candidate edges using precise mathematical criteria, substituting the approximate mechanical process with a rigorous geometric computation that preserves data integrity.

Inventive Principle:
Principle #28Mechanics substitution (Replace mechanical system)

Solution Approach 2:

The patent incorporates feedback mechanisms where the algorithm continuously evaluates the geometric properties of candidate edges and uses this information to make informed selection decisions. The feedback from geometric evaluations (such as angle measurements, edge lengths, and region containment checks) guides the edge selection process to maintain both accuracy and efficiency.

Inventive Principle:
Principle #23Feedback

3Loss of substance

If vertex-minimized contours are generated, then compression efficiency is improved, but contour accuracy may deteriorate

Engineering Contradiction:
Improvecompression efficiencyVSAvoidcontour accuracy
Core Design Contradiction:
Loss of substanceVSManufacturing precision

Solution Approach 1:

The patent performs preliminary geometric evaluations of candidate edges before finalizing the contour. By pre-evaluating the geometric properties and selecting edges that satisfy specific criteria, the algorithm ensures that the resulting vertex-minimized contour maintains high accuracy. This preliminary action prevents the need for subsequent corrections and preserves contour fidelity.

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The patent segments the contour generation process into distinct stages: identifying candidate edges, evaluating their geometric properties, selecting optimal edges, and constructing the final contour. This segmentation allows each stage to be optimized independently, ensuring that vertex minimization does not compromise overall contour accuracy while maintaining compression efficiency.

Inventive Principle:
Principle #1Segmentation

Data Source

PatentUS20230401720A1Systems and methods for contouring
Publication Date: 2023.12.14 BOARD OF RGT THE UNIV OF TEXAS SYST
  • US20230401720A1 patent drawing
  • US20230401720A1 patent drawing
  • US20230401720A1 patent drawing

AI summary

Systems and methods for image processing generate a contour for a region in an image. One example method generally includes identifying a first contour edge associated with a region having a plurality of region points, the first contour edge extending from a first region point of the plurality of region points to a second region point of the plurality of region points, identifying a second contour edge associated with the region extending from the first region point to a third region point of the plurality of region points, determining whether a contour associated with the region has fewer vertices using the second contour edge as compared to using the first contour edge, the contour comprising a concave hull, and generating the contour based on the determination.