This article presents: (i) a multi-scale representation of grey-level shape called the scale-space primal sketch, which makes explicit both features in scale-space and the relations between structures at different scales, (ii) a methodology for extracting significant blob-like image structures from this representations, and (iii) applications to edge detection, histogram analysis, and junction classification demonstrating how the proposed method can be used for guiding later stage visual processes. The representation gives a qualitative description of image structure, which allows for detection of stable scales and associated regions of interest in a solely bottom-up data-driven way. In other words, it generates coarse segmentation cues, and can hence be seen as preceding further processing, which can then be properly tuned. It is argued that once such information is available, many other processing tasks can become much simpler. Experiments on real imagery demonstrate that the proposed theory gives intuitive results.

We introduce a novel view-based object representation, called the saliency map graph (SMG), which captures the salient regions of an object view at multiple scales using a wavelet transform. This compact representation is highly invariant to translation, rotation (image and depth), and scaling, and offers the locality of representation required for occluded object recognition. To compare two saliency map graphs, we introduce two graph similarity algorithms. The first computes the topological similarity between two SMG's, providing a coarse-level matching of two graphs. The second computes the geometrical similarity between two SMG's, providing a fine-level matching of two graphs. We test and compare these two algorithms on a large database of model object views. Keywords: View-Based Object Recognition, Shape Representation and Recovery, Graph Matching. 1 Introduction The view-based approach to 3-D object recognition represents an object as a collection of 2-D views, sometimes called...

This paper explores the use of hierarchical object representations in terms of multiscale image features for simultaneous tracking and recognition of objects. Specifically, we consider an application to hand gesture analysis, where hand models are tracked over multiple postures (states). We propose a scale-invariant dissimilarity measure for comparing scale-space features. Based on it, we evaluate the likelihood of hierarchical, parameterized models containing di erent types of image features at multiple scales. The likelihood is constructed in such a way, that its maximization over different models and their parameters allows for both model selection and parameter estimation. These ideas are integrated with the framework of particle filtering, involving simultaneous tracking and recognition, and where a coarse-to-fine evaluation strategy improves computational efficiency. Based on the proposed approach, an application DrawBoard is developed, where the user controls a drawing device with a set of qualitative hand states and quantitative hand motions.

A novel approach to pattern matching is presented, which reduces time complexity by two orders of magnitude compared to traditional approaches. The suggested approach uses an efficient projection scheme which bounds the distance between a pattern and an image window using very few operations. The projection framework is combined with a rejection scheme which allows rapid rejection of image windows that are distant from the pattern. Experiments show that the approach is effective even under very noisy conditions. The approach described here can also be used in classification schemes where the projection values serve as input features that are informative and fast to extract. 1.

This paper shows how the performance of feature trackers can be improved by building a hierarchical view-based object representation consisting of qualitative relations between image structures at di#erent scales. The idea is to track all image features individually, and to use the qualitative feature relations for avoiding mismatches, resolving ambiguous matches and for introducing feature hypotheses whenever image features are lost. Compared to more traditional work on view-based object tracking, this methodology has the ability to handle semi-rigid objects and partial occlusions. Compared to trackers based on threedimensional object models, this approach is much simpler and of a more generic nature. A hands-on example is presented showing how an integrated application system can be constructed from conceptually very simple operations.

Normalized cross correlation has been used extensively for many machine vision applications, but the traditional normalized correlation operation does not meet speed requirements for time-critical applications. In this paper, we propose a fast normalized cross correlation computation for defect detection application. A sum-table scheme is utilized, which allows the calculations of image mean, image variance and cross-correlation between images to be invariant to the size of template window. Given larger images of size M × N and the neighborhood window of size m × n, the computational complexity can be significantly reduced from O ( m ⋅ n ⋅ M ⋅ N) with the traditional normalized correlation operation to only O ( M ⋅ N) with the proposed sum-table scheme.

The retrieval of images from a large database of images is an important and emerging area of research. Here, a technique to retrieve images based on appearance that works effectively across large changes of scale is proposed. The database is initially filtered with derivatives of a Gaussian at several scales. A user defined template is then created from an image of an object similar to those being sought. The template is also filtered using Gaussian derivatives. The template is then matched with the filter outputs of the database images and the matches ranked according to the match score. Experiments demonstrate the technique on a number of images in a database. No prior segmentation of the images is required and the technique works with viewpoint changes up to 20 degrees and illumination changes. 1 Introduction The advent of multi-media and large image collections in several different domains brings forth a necessity for image retrieval systems. These systems will This work was s...

This paper presents algorithms and a prototype system for hand tracking and hand posture recognition. Hand postures are represented in terms of hierarchies of multi-scale colour image features at different scales, with qualitative inter-relations in terms of scale, position and orientation. In each image, detection of multi-scale colour features is performed. Hand states are then simultaneously detected and tracked using particle filtering, with an extension of layered sampling referred to as hierarchical layered sampling. Experiments are presented showing that the performance of the system is substantially improved by performing feature detection in colour space and including a prior with respect to skin colour. These components have been integrated into a real-time prototype system, applied to a test problem of controlling consumer electronics using hand gestures. In a simplified demo scenario, this system has been successfully tested by participants at two fairs during 2001.

This paper draws on results from Computational Geometry, Differential Geometry, and recent theories on shape representation by the Medial Axis Transform (MAT) in order to introduce a novel technique of multi-scale skeletonization. To this end first the continuous medial axis transform is approximated by the Voronoi tessellation (VT) of the boundary points of a shape. Mapping features of the object boundary evolving under the heat equation onto the VT establishes the so-called skeleton-space, a shape description which combines contour characteristics with region information furnished by the tessellation. The skeleton-space is capable of describing complex shapes characterized by significantly jagged boundaries. In addition, tracking the evolution of characteristic loci of the MAT such as skeleton nodes over varying scale permits to assess the most significant skeleton constituents and to automatically determine feasible scale and pruning parameters. Submitted for publication in the IEEE...

The retrieval of images from a large database of images is an important area of current research. Here, a technique to retrieve images based on appearance is proposed. The database is initially filtered with derivatives of a Gaussian at several scales. A user defined template is then created from an image of an object similar to those being sought. The template is also filtered using Gaussian derivatives. The template is then matched with the filter outputs of the database images and the matches ranked according to the match score. Experiments demonstrate the technique on a number of images in a database. No prior segmentation of the images is required and the technique works with viewpoint changes up to 20 degrees and illumination changes. Keywords: Filter-based representations, Appearance-based representations, Scale-space Matching, Vector Correlation, Image Retrieval, Applications, Shape Representation. This work was supported in part by the Center for Intelligent Information Retr...