PointSSIM: A novel low dimensional resolution invariant image-to-image comparison metric

2025-06-30

AI Summary

To address the challenge of robustly comparing binary images across varying resolutions, this paper proposes PointSSIM—a low-dimensional, resolution-invariant image similarity metric. The method converts binary images into marked point patterns and identifies local adaptive extremal points via minimum-distance transform as structural anchors. It then constructs a structure-aware, low-dimensional summary vector by integrating intensity, connectivity, complexity, and structural similarity features. Crucially, PointSSIM eschews pixel-level alignment and instead achieves scale-invariance through morphology-guided anchor modeling. Experimental results demonstrate that PointSSIM maintains high stability and discriminative power across multiple resolutions, significantly outperforming conventional SSIM as well as histogram-based and feature-matching approaches. Its robustness and efficiency make it particularly suitable for multi-scale image assessment tasks in medical imaging and remote sensing analysis.

Technology Category

• Computer Vision: Image and Video Retrieval
• Machine Learning: Large Multimodal Models (LMMs)
• Search and Optimization: Sampling/Simulation-based Search

Application Category

• Security and Privacy: Large-scale security measurements
• Search and Retrieval-Augmented AI: Web evaluation methodologies and metrics
• Web Mining and Content Analysis: Robustness and generalizability of Web mining methods

Abstract

This paper presents PointSSIM, a novel low-dimensional image-to-image comparison metric that is resolution invariant. Drawing inspiration from the structural similarity index measure and mathematical morphology, PointSSIM enables robust comparison across binary images of varying resolutions by transforming them into marked point pattern representations. The key features of the image, referred to as anchor points, are extracted from binary images by identifying locally adaptive maxima from the minimal distance transform. Image comparisons are then performed using a summary vector, capturing intensity, connectivity, complexity, and structural attributes. Results show that this approach provides an efficient and reliable method for image comparison, particularly suited to applications requiring structural analysis across different resolutions.

Problem

• Develops resolution-invariant image comparison metric
• Extracts key features from binary images
• Enables structural analysis across varying resolutions

Innovation

• Resolution invariant image comparison metric
• Marked point pattern representations
• Summary vector for structural attributes