From line segments to more organized Gestalts

TL;DR

This paper introduces an unsupervised, bottom-up computer vision method that detects higher-level Gestalt features like good continuations, alignments, and bars from elementary line segments using a stochastic a contrario model.

Contribution

It proposes a novel unsupervised algorithm that builds higher-level geometric structures from line segments based on Gestalt principles, advancing bottom-up image analysis.

Findings

Successfully detects Gestalt features in digital images

Uses a stochastic a contrario model for reliable detection

Provides formulas with controlled false alarm rates

Abstract

In this paper, we reconsider the early computer vision bottom-up program, according to which higher level features (geometric structures) in an image could be built up recursively from elementary features by simple grouping principles coming from Gestalt theory. Taking advantage of the (recent) advances in reliable line segment detectors, we propose three feature detectors that constitute one step up in this bottom up pyramid. For any digital image, our unsupervised algorithm computes three classic Gestalts from the set of predetected line segments: good continuations, nonlocal alignments, and bars. The methodology is based on a common stochastic {\it a contrario model} yielding three simple detection formulas, characterized by their number of false alarms. This detection algorithm is illustrated on several digital images.