scispace - formally typeset
Search or ask a question
Proceedings ArticleDOI

Hierarchical Grouping Using Gestalt Assessments

01 Oct 2017-pp 1702-1709
TL;DR: The work in this paper starts from super-pixel primitives, and the grouping ends when the Gestalten almost fill the whole image, when the recognition rates are a little better.
Abstract: Real images contain symmetric Gestalten with high probability. I.e. certain parts can be mapped on other certain parts by the usual Gestalt laws and are repeated there with high similarity. Moreover, such mapping comes in nested hierarchies - e.g. a reflection Gestalt that is made of repetition friezes, whose parts are again reflection symmetric compositions. This can be explicitly modelled by continuous assessment functions. Hard decisions on whether or not a law is fulfilled are avoided. Starting from primitive objects extracted from the input image successively aggregates are constructed: reflection pairs, rows, etc., forming a part-of-hierarchy and rising in scale. The work in this paper starts from super-pixel primitives, and the grouping ends when the Gestalten almost fill the whole image. Occasionally the results may not be in accordance with human perception. The parameters have not been adjusted specifically for the data at hand. Previous work only used the compulsory attributes location, scale, orientation and assessment for each object. A way to improve the recognition performance is utilizing additional features such as colors or eccentricity. Thus the recognition rates are a little better.

Content maybe subject to copyright    Report

Citations
More filters
Proceedings ArticleDOI
01 Jul 2017
TL;DR: This report provides a detailed summary of the evaluation methodology for each type of symmetry detection algorithm validated, and demonstrates and analyzes quantified detection results in terms of precision-recall curves and F-measures for all algorithms evaluated.
Abstract: Motivated by various new applications of computational symmetry in computer vision and in an effort to advance machine perception of symmetry in the wild, we organize the third international symmetry detection challenge at ICCV 2017, after the CVPR 2011/2013 symmetry detection competitions. Our goal is to gauge the progress in computational symmetry with continuous benchmarking of both new algorithms and datasets, as well as more polished validation methodology. Different from previous years, this time we expand our training/testing data sets to include 3D data, and establish the most comprehensive and largest annotated datasets for symmetry detection to date; we also expand the types of symmetries to include densely-distributed and medial-axis-like symmetries; furthermore, we establish a challenge-and-paper dual track mechanism where both algorithms and articles on symmetry-related research are solicited. In this report, we provide a detailed summary of our evaluation methodology for each type of symmetry detection algorithm validated. We demonstrate and analyze quantified detection results in terms of precision-recall curves and F-measures for all algorithms evaluated. We also offer a short survey of the paper-track submissions accepted for our 2017 symmetry challenge.

44 citations


Cites background or methods or result from "Hierarchical Grouping Using Gestalt..."

  • ...In “Hierarchical Grouping Using Gestalt Assessments” [30], Michaelsen and Arens describe a framework for using various types of symmetry (e....

    [...]

  • ...The algorithms are the challenger Michaelsen and Arens [30] (MA), Funk and Liu [14] (Sym-VGG and SymRes - the baselines), Loy and Eklundh [26] (LE), Tsogkas and Kokkinos [47] (MIL), Teo et al....

    [...]

  • ...For 1D translation symmetry, the baseline outperformed the submission of Michaelsen and Arens [30] (Figure 5)....

    [...]

  • ...The algorithms are Michaelsen and Arens [30], Elawady et al....

    [...]

  • ...[41, 42] (LMSDS,FSDS), and the challenge submission from Michaelsen and Arens [30]....

    [...]

Journal ArticleDOI
TL;DR: A new approach based on the Slope Chain Code to determine whether a curve is rotational symmetrical and its order of symmetry is presented, which works for open and closed perfectly symmetrical or quasi-symmetrical 2D curves.

9 citations

Journal ArticleDOI
TL;DR: In this article, a stable metric is proposed to extract subsets of consistently oriented candidate segments, whenever the underlying 2D signal appearance exhibits definite near symmetric correspondences, and the ranking of such segments on the basis of the surrounding gradient orientation specularity, in order to reflect real symmetric object boundaries.
Abstract: This work addresses the challenging problem of reflection symmetry detection in unconstrained environments. Starting from the understanding on how the visual cortex manages planar symmetry detection, it is proposed to treat the problem in two stages: i) the design of a stable metric that extracts subsets of consistently oriented candidate segments, whenever the underlying 2D signal appearance exhibits definite near symmetric correspondences; ii) the ranking of such segments on the basis of the surrounding gradient orientation specularity, in order to reflect real symmetric object boundaries. Since these operations are related to the way the human brain performs planar symmetry detection, a better correspondence can be established between the outcomes of the proposed algorithm and a human-constructed ground truth. When compared to the testing sets used in recent symmetry detection competitions, a remarkable performance gain can be observed. In additional, further validation has been achieved by conducting perceptual validation experiments with users on a newly built dataset.

6 citations

Proceedings ArticleDOI
12 May 2019
TL;DR: This work exploits the estimated boundary of the object and describes a boundary pixel using only the estimated normal of the boundary segment around the pixel to embed the symmetry axes in a graph as cliques to robustly detect the symmetry axis.
Abstract: Reflection symmetry is ubiquitous in nature and plays an important role in object detection and recognition tasks. Most of the existing methods for symmetry detection extract and describe each keypoint using a descriptor and a mirrored descriptor. Two keypoints are said to be mirror symmetric key-points if the original descriptor of one keypoint and the mirrored descriptor of the other keypoint are similar. However, these methods suffer from the following issue. The background pixels around the mirror symmetric pixels lying on the boundary of an object can be different. Therefore, their descriptors can be different. However, the boundary of a symmetric object is a major component of global reflection symmetry. We exploit the estimated boundary of the object and describe a boundary pixel using only the estimated normal of the boundary segment around the pixel. We embed the symmetry axes in a graph as cliques to robustly detect the symmetry axes. We show that this approach achieves state-of-the-art results in a standard dataset.

3 citations


Cites background from "Hierarchical Grouping Using Gestalt..."

  • ...The approaches for reflection symmetry detection in real-world images can be categorized as voting based approaches [12, 13, 14, 21, 4, 6, 5, 15, 16, 17] and multiple model fitting approaches in [18, 19, 20]....

    [...]

Book ChapterDOI
01 Jan 2019
TL;DR: This chapter is meant as an attempt to code machines so that they become better, understanding what used to be a “pattern language” as the closure of the Gestalt Algebra and presenting ways to propagate the GestAlt laws through the hierarchy, if necessary, down to the preceding primitives.
Abstract: Gestalt perception comes in hierarchies. Human observers are quick and reliable in reconstructing from pictorial data constructions like a row made of reflection symmetric parts, where each wing of the parts is again a lattice of reflection symmetric sub-parts, etc. Machines hardly compete with these skills up to now, in particular when projection distorted the patterns, or occlusion deleted some parts of it. This chapter is meant as an attempt to code machines so that they become better. To this end, the combinatorial nature of such “Gestalt sentences” must first be understood. Instead of the traditional way of defining grammars, here an algebraic view is taken, understanding what used to be a “pattern language” as the closure of the Gestalt Algebra. The operations of this algebra are the Gestalt operations given in the previous chapters. Of course such closure is infinite, but we can prove that almost all Gestalten in it have assessments close to zero. The set of hierarchical Gestalten assessed better than an \(\epsilon >0\) and resulting from a finite set of primitives is finite. Still, this set can be very huge. In particular if we use the operations simply, i.e., only taking into account the features of immediately preceding Gestalten. However, in this chapter we also present ways to propagate the Gestalt laws through the hierarchy, if necessary, down to the preceding primitives. This greatly reduces numbers and efforts. The literature on symmetry recognition sees a problem in the non-local nature of, e.g., reflection symmetry. Search for correspondence cannot be bounded by less than quadratic complexity. Here hierarchy can actually help. Correspondence between very distant small objects is established by use of hierarchy. This can be implemented in sub-quadratic complexity. Thus, what looks at first glance like a combinatorial nightmare, turns out to be a proposal for the solution of an old and hard combinatorial correspondence problem.
References
More filters
Journal ArticleDOI
TL;DR: A new superpixel algorithm is introduced, simple linear iterative clustering (SLIC), which adapts a k-means clustering approach to efficiently generate superpixels and is faster and more memory efficient, improves segmentation performance, and is straightforward to extend to supervoxel generation.
Abstract: Computer vision applications have come to rely increasingly on superpixels in recent years, but it is not always clear what constitutes a good superpixel algorithm. In an effort to understand the benefits and drawbacks of existing methods, we empirically compare five state-of-the-art superpixel algorithms for their ability to adhere to image boundaries, speed, memory efficiency, and their impact on segmentation performance. We then introduce a new superpixel algorithm, simple linear iterative clustering (SLIC), which adapts a k-means clustering approach to efficiently generate superpixels. Despite its simplicity, SLIC adheres to boundaries as well as or better than previous methods. At the same time, it is faster and more memory efficient, improves segmentation performance, and is straightforward to extend to supervoxel generation.

7,849 citations

Book
29 Oct 1993
TL;DR: This book presents a meta-modelling framework for analysing two or more samples of unimodal data from von Mises distributions, and some modern Statistical Techniques for Testing and Estimation used in this study.
Abstract: Preface 1. The purpose of the book 2. Survey of contents 3. How to use the book 4. Notation, terminology and conventions 5. Acknowledgements Part I. Introduction: Part II. Descriptive Methods: 2.1. Introduction 2.2. Data display 2.3. Simple summary quantities 2.4. Modifications for axial data Part III. Models: 3.1. Introduction 3.2. Notation trigonometric moments 3.3. Probability distributions on the circle Part IV. Analysis of a Single Sample of Data: 4.1. Introduction 4.2. Exploratory analysis 4.3. Testing a sample of unit vectors for uniformity 4.4. Nonparametric methods for unimodal data 4.5. Statistical analysis of a random sample of unit vectors from a von Mises distribution 4.6. Statistical analysis of a random sample of unit vectors from a multimodal distribution 4.7. Other topics Part V. Analysis of Two or More Samples, and of Other Experimental Layouts: 5.1. Introduction 5.2. Exploratory analysis 5.3. Nonparametric methods for analysing two or more samples of unimodal data 5.4. Analysis of two or more samples from von Mises distributions 5.5. Analysis of data from more complicated experimental designs Part VI. Correlation and Regression: 6.1. Introduction 6.2. Linear-circular association and circular-linear association 6.3. Circular-circular association 6.4. Regression models for a circular response variable Part VII. Analysis of Data with Temporal or Spatial Structure: 7.1. Introduction 7.2. Analysis of temporal data 7.3. Spatial analysis Part VIII. Some Modern Statistical Techniques for Testing and Estimation: 8.1. Introduction 8.2. Bootstrap methods for confidence intervals and hypothesis tests: general description 8.3. Bootstrap methods for circular data: confidence regions for the mean direction 8.4. Bootstrap methods for circular data: hypothesis tests for mean directions 8.5. Randomisation, or permutation, tests Appendix A. Tables Appendix B. Data sets References Index.

2,323 citations


Additional excerpts

  • ...g, Riess distributions can be used [16]....

    [...]

Journal ArticleDOI
01 Jan 1923

1,905 citations

Journal ArticleDOI

827 citations


Additional excerpts

  • ...g, Riess distributions can be used [16]....

    [...]

Book ChapterDOI
07 May 2006
TL;DR: It is shown how symmetric pairs of features can be efficiently detected, how the symmetry bonding each pair is extracted and evaluated, and how these can be grouped into symmetric constellations that specify the dominant symmetries present in the image.
Abstract: A novel and efficient method is presented for grouping feature points on the basis of their underlying symmetry and characterising the symmetries present in an image. We show how symmetric pairs of features can be efficiently detected, how the symmetry bonding each pair is extracted and evaluated, and how these can be grouped into symmetric constellations that specify the dominant symmetries present in the image. Symmetries over all orientations and radii are considered simultaneously, and the method is able to detect local or global symmetries, locate symmetric figures in complex backgrounds, detect bilateral or rotational symmetry, and detect multiple incidences of symmetry.

387 citations


"Hierarchical Grouping Using Gestalt..." refers background or methods in this paper

  • ...of certain primitive objects extracted from the image – like in [1], or to fill certain accumulators directly from the raw colors – like in Hough transform methods....

    [...]

  • ...Including SIFT 128-dimensional key-point features in order to improve the performance following [1] was demonstrated in [10]....

    [...]

  • ...References [1] G. Loy and J.-O. Eklundh: Detecting symmetry and symmetric constellations of features....

    [...]

  • ...The reason might have been, that SIFT points provide exactly the desired Gestalt domain features location, scale, orientation, and assessment, and maybe also because the standard solution of Loy & Eklundh [1] also was based on these....

    [...]