Oliver Tiebe

Object matching with hierarchical skeletons

The skeleton of an object provides an intuitive and effective abstraction which facilitates object matching and recognition. However, without any human interaction, traditional skeleton-based descriptors and matching algorithms are not stable for deformable objects. Specifically, some fine-grained topological and geometrical features would be discarded if the skeleton was incomplete or only represented significant visual parts of an object. Moreover, the performance of skeleton-based matching highly depends on the quality and completeness of skeletons. In this paper, we propose a novel object representation and matching algorithm based on hierarchical skeletons which capture the shape topology and geometry through multiple levels of skeletons. For object representation, we reuse the pruned skeleton branches to represent the coarse- and fine-grained shape topological and geometrical features. Moreover, this can improve the stability of skeleton pruning without human interaction. We also propose an object matching method which considers both global shape properties and fine-grained deformations by defining singleton and pairwise potentials for similarity computation between hierarchical skeletons. Our experiments attest our hierarchical skeleton-based method a significantly better performance than most existing shape-based object matching methods on six datasets, achieving a 99.21% bulls-eye score on the MPEG7 shape dataset. HighlightsIt represents the coarse- and fine-grained shape topological and geometrical features.It improves the stability of skeleton pruning without human interaction.It considers both global and fine shape properties by different potential functions.It achieves a better performance than most existing methods on six datasets.Experiments attest our method a better performance than most related approaches.We achieve a 99.21% bulls-eye score on the MPEG7 shape dataset.

Shape-based object retrieval by contour segment matching

In this paper we introduce an approach for object retrieval that uses contour segment matching for shape similarity computation. The object contour is partitioned into segments by skeleton endpoints. Each contour segment is represented by a rotation and scale invariant, 12-dimensional feature vector. The similarity of two objects is determined by matching their contour segments using the Hungarian algorithm. Our method is insensitive to object deformation and outperforms existing shape-based object retrieval algorithms. The most significant scientific contributions of this paper include (i) the introduction of a new feature extraction technique for contour segments as well as (ii) a new similarity measure for contour segments cleverly modelling the human perception and easily adapting to concrete application domains, and (iii) the impressive robustness of the method in an object retrieval scenario.

Shape-Based Classification of Environmental Microorganisms

Occurrence of certain environmental microorganisms and their species is a very informative indicator to evaluate environmental quality. Unfortunately, their manual recognition in microbiological laboratories is very time-consuming and expensive. Therefore, we work on an automatic method for shape-based classification of EMs in microscopic images. First, we segment the microorganisms from the background. Second, we describe their shapes by discriminative feature vectors. Third, we perform the EM classification using Support Vector Machines. The most important scientific contribution of this paper, in comparison to the state-of-the-art and to our previous publications in this field, is the introduction of a completely new and very robust 2D feature descriptor for EM shapes. Experimental results certify the effectiveness and practicability of our automatic EM classification system emphasising the benefits achieved with the new shape descriptor proposed in this work.

Shape-based Object Matching Using Point Context

This paper proposes a novel object matching algorithm based on shape contours. In order to ensure low computational complexity in shape representation, our descriptor is composed by a small number of interest points which are generated by considering both curvatures and the overall shape trend. To effectively describe each point of interest, we introduce a simple and highly discriminative point descriptor, namely Point Context, which represents its geometrical and topological location. For shape matching, we observed that the correspondences are not only dependent on the similarities between these single points in different objects, but they are also related to the geometric relations between multiple points of interest in the same object. Therefore, a high-order graph matching formulation is introduced to merge the single point similarities and the similarities between point triangles. The main contributions of this paper include (i) the introduction of a novel shape descriptor with robust shape points and their descriptors and (ii) the implementation of a high-order graph matching algorithm that solves the shape matching problem. Our method is validated through a series of object retrieval experiments on four datasets demonstrating its robustness and accuracy.

Shape-based object matching using interesting points and high-order graphs

The introduction of a novel shape descriptor with robust shape interesting points and their descriptors.The implementation of a high-order graph matching algorithm for solving the shape matching problem.The proposed method can significantly improve the traditional correspondence-based shape matching methods.The proposed method is very robust in an object retrieval scenario. In shape-based object matching, it is important how to fuse similarities between points on a shape contour and the ones on another contour into the overall similarity. However, existing methods face two critical problems. Firstly, since most contour points are involved for possible matchings without taking into account the usefulness of each point, it causes high computational costs for point matching. Secondly, existing methods do not consider geometrical relations characterised by multiple points. In this paper, we propose a shape-based object matching method which is able to overcome these problems. To counteract the first problem mentioned, we devise a shape descriptor using a small number of interesting points which are generated by considering both curvatures and the overall shape trend. We also introduce a simple and highly discriminative point descriptor, namely Point Context, which represents the geometrical and topological location of each interesting point. For the second problem, we employ high-order graph matching which examines similarities for singleton, pairwise and triple relations of points. We validate the robustness and accuracy of our method through a series of experiments on six datasets.

Shape-based Object Retrieval and Classification with Supervised Optimisation

In order to enhance the performance of shape retrieval and classification, in this paper, we propose a novel shape descriptor with low computation complexity that can be easily fused with other meaningful descriptors like shape context, etc. This leads to a significant increase in descriptive power of original descriptors without adding to much computation complexity. To make the proposed shape descriptor more practical and general, a supervised optimisation strategy is introduced. The most significant scientific contributions of this paper includes the introduction of a new and simple feature descriptor with supervised optimisation strategy leading to the impressive improvement of the accuracy in object classification and retrieval scenario.

Shape Matching Using Point Context and Contour Segments

This paper proposes a novel method to generate robust contour partition points and applies them to produce point context and contour segment features for shape matching. The main idea is to match object shapes by matching contour partition points and contour segments. In contrast to typical shape context method, we do not consider the topological graph structure since our approach is only considering a small number of partition points rather than full contour points. The experimental results demonstrate that our method is able to produce correct results in the presence of articulations, stretching, and contour deformations. The most significant scientific contributions of this paper include (i) the introduction of a novel partition point extraction technique for point context and contour segments as well as (ii) a new fused similarity measure for object matching and recognition, and (iii) the impressive robustness of the method in an object retrieval scenario as well as in a real application for environmental microorganism recognition.

Stripes-Based Object Matching

We propose a novel and fast 3D object matching framework that is able to fully utilise the geometry of objects without any object reconstruction process. Traditionally, 3D object matching methods are mostly applied based on 3D models. In order to generate accurate and proper 3D models, object reconstruction methods are used for the collected data from laser or time-of-flight sensors. Although those methods are naturally appealing, heavy computations are required for segmentation as well as transformation estimation. Moreover, some useful features could be filtered out during the reconstruction process. On the contrary, the proposed method is applied without any reconstruction process. Building on stripes generated from laser scanning lines, we represent an object by a set of stripes. To capture the full geometry, we describe each stripe by the proposed robust point context descriptor. After representing all stripes, we perform a flexible and fast matching over all collected stripes. We show that the proposed method achieves promising results on some challenging real-life objects.

Skeleton-based audio envelope shape analysis

This paper presents a novel skeleton-based audio envelope representation and matching method for audio signal analysis. We propose using amplitude envelope in the time domain to represent and calculate the similarity between audio signals. To effectively describe the shape of each envelope, we employ the skeleton descriptor, namely Audio Skeleton, to integrate both geometrical and topological envelope features. Based on Audio Skeletons, the audio envelope matching can be substituted by searching for the correspondences of skeleton endpoints. Finally, the similarity between audio envelope shapes is calculated based on their correlated skeleton matching. Our main contributions include (i) the introduction of a skeleton-based audio envelope descriptor, (ii) a simple and efficient Audio Skeleton representation method and (iii) a fast skeleton pruning and matching algorithm.

Evaluating contour segment descriptors

Contour segment (CS) is the fundamental element of partial boundaries or edges in shapes and images. So far, CS has been widely used in many applications, including object detection/matching and open curve matching. To increase the matching accuracy and efficiency, a variety of CS descriptors have been proposed. A CS descriptor is formed by a chain of boundary or edge points and is able to encode the geometric configuration of a CS. Because many different CS descriptors exist, a structured overview and quantitative evaluation are required in the context of CS matching. This paper assesses 27 CS descriptors in a structured way. Firstly, the analytical invariance properties of CS descriptors are explored with respect to scaling, rotation and transformation. Secondly, their distinctiveness is evaluated experimentally on three datasets. Lastly, their computation complexity is studied. Based on results, we find that both CS lengths and matching algorithms affect the CS matching performance while matching algorithms have higher affection. The results also reveal that, with different combinations of CS descriptors and matching algorithms, several requirements in terms of matching speed and accuracy can be fulfilled. Furthermore, a proper combination of CS descriptors can improve the matching accuracy over the individuals.