Lech Szumilas

Object localization based on Markov random fields and symmetry interest points

We present an approach to detect anatomical structures by configurations of interest points, from a single example image. The representation of the configuration is based on Markov Random Fields, and the detection is performed in a single iteration by the MAX-SUM algorithm. Instead of sequentially matching pairs of interest points, the method takes the entire set of points, their local descriptors and the spatial configuration into account to find an optimal mapping of modeled object to target image. The image information is captured by symmetry-based interest points and local descriptors derived from Gradient Vector Flow. Experimental results are reported for two data-sets showing the applicability to complex medical data.

Invariant Shape Matching for Detection of Semi-local Image Structures

Shape features applied to object recognition has been actively studied since the beginning of the field in 1950s and remain a viable alternative to appearance based methods e.g. local descriptors. This work address the problem of learning and detecting repeatable shape structures in images that may be incomplete, contain noise and/or clutter as well as vary in scale and orientation. A new approach is proposed where invariance to image transformations is obtained through invariant matching rather than typical invariant features. This philosophy is especially applicable to shape features such as open edges which do not have a specific scale or specific orientation until assembled into an object. Our primary contributions are: a new shape-based image descriptor that encodes a spatial configuration of edge parts, a technique for matching descriptors that is rotation and scale invariant and shape clustering that can extract frequently appearing image structures from training images without a supervision.

Local Structure Detection with Orientation-invariant Radial Configuration

Local image descriptors have proved themselves as useful tools for many computer vision tasks such as matching points between multiple images of a scene and object recognition. Current descriptors, such as SIFT, are designed to match image features with unique local neighborhoods. However, the interest point detectors used with SIFT often fail to select perceptible local structures in the image, and the SIFT descriptor does not directly encode the local neighborhood shape. In this paper we propose a symmetry based interest point detector and radial local structure descriptor which consistently captures the majority of basic local image structures and provides a geometrical description of the structure boundaries. This approach concentrates on the extraction of shape properties in image patches, which are an intuitive way to represent local appearance for matching and classification. We explore the specificity and sensitivity of this local descriptor in the context of classification of natural patterns. The implications of the performance comparison with standard approaches like SIFT are discussed.

Evaluation of Gradient Vector Flow for Interest Point Detection

We present and evaluate an approach for finding local interest points in images based on the non-minima suppression of Gradient Vector Flow (GVF) magnitude. Based on the GVFs properties it provides the approximate centers of blob-like structures or homogeneous structures confined by gradients of similar magnitude. It results in a scale and orientation invariant interest point detector, which is highly stable against noise and blur. These interest points outperform the state of the art detectors in various respects. We show that our approach gives a dense and repeatable distribution of locations that are robust against affine transformations while they outperform state of the art techniques in robustness against lighting changes, noise, rotation and scale changes. Extensive evaluation is carried out using the Mikolajcyzk framework for interest point detector evaluation.

Boosted Edge Orientation Histograms for Grasping Point Detection

In this paper, we describe a novel algorithm for the detection of grasping points in images of previously unseen objects. A basic building block of our approach is the use of a newly devised descriptor, representing semi-local grasping point shape by the use edge orientation histograms. Combined with boosting, our method learns discriminative grasp point models for new objects from a set of annotated real-world images. The method has been extensively evaluated on challenging images of real scenes, exhibiting largely varying characteristics concerning illumination conditions, scene complexity, and viewpoint. Our experiments show that the method works in a stable manner and that its performance compares favorably to the state-of-the-art.

Combining Geometry and Local Appearance for Object Detection

In this paper we address the problem of object detection in cluttered scenes. Local image features and their spatial configuration act as representation of object classes which are learned in a discriminative fashion. Recent contributions in the area of object detection indicate the importance of using geometrical properties for representing object classes. Prompted by this, we devised an approach tailored to control the importance of the features and their spatial alignment. We quantitatively show that modeling the spatial distribution of local features and optimising the in???uence of both cues significantly boosts object detection performance.

Image-Based grasping point detection using boosted histograms of oriented gradients

In this paper, we describe the components of a novel algorithm for the detection of grasping points from monocular images of previously unseen objects. A basic building block of our approach is the use of a newly devised descriptor, capable of representing grasping point shape and appearance by the use of histograms of oriented gradients in a semi-local manner. Combined with boosting our method learns discriminative grasp point models for new objects from a set of annotated real-world images. The method has been extensively evaluated on challenging images of real scenes, exhibiting largely varying characteristics concerning illumination conditions, scene complexity, and viewpoint. Our experiments show that the method, despite these variations, works in a stable manner and that its performance compares favorably to the state-of-the-art.

Radial edge configuration for semi-local image structure description

We present a novel semi-local image descriptor which encodes multiple edges corresponding to the image structure boundaries around an interest point. The proposed method addresses the problem of poor edge detection through a robust, scale and orientation invariant, descriptor distance. In addition, a clustering of descriptors capable of extracting distinctive shapes from a set of descriptors is described. The proposed techniques are applied to the description of bone shapes in medical X-ray images and the experimental results are presented.

Color pair clustering for texture detection

A novel approach to the extraction of image regions of uniform color and its application to automatic texture detection is discussed. The method searches for alternating color patterns, through hierarchical clustering of color pairs from adjacent image regions. The final result is a hierarchy of texture regions, described by their boundaries and a set of features, detected at multiple accuracy levels. The results are presented on some images of natural scenes from the Berkeley segmentation dataset and benchmark.