Andrzej Sluzek

Identification and inspection of 2-D objects using new moment-based shape descriptors

A new method of using the moment invariants for the identification and inspection of 2-D shapes is proposed. The prototype object is described by a family of shapes. These shapes are created by occluding the object by circles (of different radius) located in the objects centre of area. The moment invariants of such shapes are functions of a parameter describing the size of circles. Using these functions it is possible to create from a single moment invariant many shape descriptors, so there is no need to use the moments of higher order. The application of the method to the quality inspection of industrial parts is presented.

On moment-based local operators for detecting image patterns

Abstract Local operators, template matching and moments are widely used in image processing. In this paper, they are combined into an efficient method of designing detectors for various image patterns. The method is using a circular window of the size corresponding to the problem requirements. For each location of the window, moment features are used to determine the optimum template that is subsequently matched to the actual content of the window to produce a ‘pattern intensity’ map. Theory and general recommendations are illustrated by results obtained for exemplary patterns. The method is also briefly compared to other techniques.

A low-dimensional local descriptor incorporating TPS warping for image matching

This paper proposes a low-dimensional image descriptor combining shape characteristics and location information. The shape characteristics are obtained from simple rectangular patterns approximating the interest regions. Although the shape component of the descriptor does not perform as well as high-dimensional descriptors (e.g. SIFT) it can be built very quickly. Moreover, it usually provides enough correspondences to align locations of interest regions. The alignment is based on the thin plate spline (TPS) warping algorithm with control points automatically identified by our method. Subsequently, the aligned coordinates contribute additional dimensions to the descriptor. The process may be iterated several times until no further improvement is achieved. Experiments show that incorporation of location data into the descriptor improves performance. The proposed descriptor is compared to SIFT (a standard benchmark which is considered one of the best local descriptors [1]) for real images with various geometric and photometric transformations and for diversified types of scenes. Results show the proposed low-dimensional descriptor generally performs better than SIFT descriptor while the computational complexity of our descriptor is far superior.

Model of gated imaging in turbid media

Range-gated imaging can improve the signal to backscattering noise ratio (SBR) in turbid media. This is achieved by synchronizing a short duration, high intensity pulse with precise camera gating. It is well known that shorter pulse length and shorter camera gate duration can enhance the SBR. However, there is no analytical model of the backscattering noise (as a function of the pulse length and gate timing) that can be used to minimize backscattering noise within the camera-captured signal. We propose a formulation (a modification of Falks lidar equation) that models the backscattering noise as a convolution with a fixed upper limit. This formulation predicts a variation of backscattering noise within the returning signal. In particular, the model predicts higher SBR toward the tail region of the target-reflected irradiance. It confirms the experimental results reported by other authors. Additionally, the model explains experimentally observed SBR improvement for shorter pulses and shorter gating intervals (if adequately positioned within the returning pulse).

Embedding Visual Words into Concept Space for Action and Scene Recognition

In this paper we propose a novel approach to introducing semantic relations into the bag-of-words framework. We use the latent semantic models, such as LSA and pLSA, in order to define semantically-rich features and embed the visual features into a semantic space. The semantic features used in LSA technique are derived from the low-rank approximation of word-document occurrence matrix by SVD. Similarly, by using the pLSA approach, the topic-specific distributions of words can be considered dimensions of a concept space. In the proposed space, the distances between words represent the semantic distances which are used for constructing a discriminative and semantically meaningful vocabulary. We have tested our approach on the KTH action database and on the Fifteen Scene database and have achieved very promising results on both.

A Survey on System-Level Techniques for Power Reduction in Field Programmable Gate Array (FPGA)-Based Devices

With growth demands to untethered embedded systems, e.g. sensor nodes, in the flexibility, performance, product longevity, areas, and decreasing time-to-market (TTM), programmable logic devices may allow covering lack of suitable processing units. Decrease in the feature transistor sizes allows producing smaller programmable devices and providing more computational power. However, such shrinking of feature sizes introduces higher power lost (static power). Moreover, programmable devices are clocked with higher frequencies due to the performance demands increase that introduces additional power consumption. Therefore, there is a need for techniques that allow for substantial power reduction and being achievable on the higher levels of the designing process. In this paper, we address means of the dynamic power consumption reduction on the system-level. This is envisaged that proposed techniques may allow achieving substantial power consumption savings with negligible hardware overheads while maintaining the energy per operation.

Relative scale method to locate an object in cluttered environment

This paper proposes an efficient method to locate a three-dimensional object in cluttered environment. Model of the object is represented in a reference scale by the local features extracted from several reference images. A PCA-based hashing technique is introduced for accessing the database of reference features efficiently. Localization is performed in an estimated relative scale. Firstly, a pair of stereo images is captured simultaneously by calibrated cameras. Then the object is identified in both images by extracting features and matching them with reference features, clustering the matched features with generalized Hough transformation, and verifying clusters with spatial relations between the features. After the identification process, knowledge-based correspondences of features belonging to the object present in the stereo images are used for the estimation of the 3D position. The localization method is robust to different kinds of geometric and photometric transformations in addition to cluttering, partial occlusions and background changes. As both the model representation and localization are single-scale processes, the method is efficient in memory usage and computing time. The proposed relative scale method has been implemented and experiments have been carried out on a set of objects. The method results very good accuracy and takes only a few seconds for object localization by our primary implementation. An application of the relative scale method for exploration of an object in cluttered environment is demonstrated. The proposed method could be useful for many other practical applications.

Image-Based Information Guide on Mobile Devices

We present a prototype of an information guide system to be used outdoor in our campus. It allows a user to find places of interest (e.g., lecture halls and libraries) using a camera phone. We use a database of panoramic views of campus scenes tagged by GPS locations, which can diminish overlapping between views. Panoramic views with the closest locations with the query view are acquired. We exploit a wide-baseline matching technique to match between views. However, due to dissimilarity in viewpoints and presence of repetitive structures, a vast percentage of matches could be false matches. We propose a verification model to effectively eliminate false matches. The true correspondences are chosen for pose recovery and information is then projected onto the image. The system is validated by extensive experiments, with images taken in different seasons, weather, illumination conditions, etc.

Development of a laser-based vision system for an underwater vehicle

The paper describes two crucial modules of a laser-based vision system for an underwater welding vehicle which is a joint research project of Nanyang Technological University (Singapore) and a major international corporation. The vehicle would automatically approach an underwater object, find weld seams, detect cracks and finally weld them. The first module is the ranging system. The experiments have shown that much better accuracy has been achieved when approximating polynomial functions are used instead of standard triangulation techniques. The second module detects and localises weld seams by extracting shape primitives corresponding to the expected configuration of seams from the laser-produced contours. Although the experiments are currently conducted in air, the proposed solution should be equally suitable for an underwater environment as well.

Visual similarity issues in face recognition

The paper discusses several issues of visual similarity in face detection and recognition. Using a straightforward concept of keypoint correspondences, a method is proposed to formalise the subjective impressions of |similar faces|, |similar eyes|, |similar chins|, etc. The method exploits the mechanism of affine near-duplicate fragment detection originally proposed for visual information retrieval. It is shown that using such a method, a simple and relatively reliable face detection/identification systems can be build without any model (or training) of human faces, which can work with images containing multiple faces shown on random backgrounds. Additionally, it is proposed how the same approach can be used to optimise databases of face images and to identify individuals who are at higher risks of mistaken face identification.