From line segments to more organized Gestalts
This work addresses the challenge of organizing low-level visual features into meaningful structures for computer vision applications, representing an incremental step in the bottom-up vision program.
The paper tackles the problem of detecting higher-level geometric structures in images by proposing an unsupervised algorithm that computes three classic Gestalt features—good continuations, nonlocal alignments, and bars—from pre-detected line segments, using a stochastic a contrario model to characterize detections by their number of false alarms.
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.