This PDF file contains the front matter associated with SPIE Proceedings Volume 8000, including the Title Page, Copyright information, Table of Contents, Introduction, and Conference Committee listing.

Accuracy of 3D vision-based reconstruction tasks depends both on the complexity of analyzed objects and on good viewing / illumination conditions, ensuring image quality and minimizing consequently measurement errors after processing of acquired images. In this contribution, as a complement to an autonomous cognitive vision system automating 3D reconstruction and using Situation Graph Trees (SGTs) as a planning / control tool, these graphs are optimized in two steps. The first (off-line) step addresses the placement of lighting sources, with the aim to find positions minimizing processing errors during the subsequent reconstruction steps. In the second step, on-line application of the SGT-based control module focuses on adjustment of illumination conditions (e. g., intensity), leading eventually to process re-planning, and enabling further to extract optimally the contour data required for 3D reconstruction. The whole illumination optimization procedure has been fully automated and included in the dynamic (re-)planning tool for visionbased reconstruction tasks, e. g. in view of quality control applications.

A fabric's smoothness is a key factor to determine the quality of textile finished products and has great influence on the functionality of industrial textiles and high-end textile products. With popularization of the 'zero defect' industrial concept, identifying and measuring defective material in the early stage of production is of great interest for the industry. In the current market, many systems are able to achieve automatic monitoring and control of fabric, paper, and nonwoven material during the entire production process, however online measurement of hairiness is still an open topic and highly desirable for industrial applications. In this paper we propose a computer vision approach, based on variable homography, which can be used to measure the emergent fiber's length on textile fabrics. The main challenges addressed in this paper are the application of variable homography to textile monitoring and measurement, as well as the accuracy of the estimated calculation. We propose that a fibrous structure can be considered as a two-layer structure and then show how variable homography can estimate the length of the fiber defects. Simulations are carried out to show the effectiveness of this method to measure the emergent fiber's length. The true lengths of selected fibers are measured precisely using a digital optical microscope, and then the same fibers are tested by our method. Our experimental results suggest that smoothness monitored by variable homography is an accurate and robust method for quality control of important industrially fabrics.

It is very important to guarantee the quality of the industrial products by means of visual inspection. In order to reduce the soldering defect with terminal deformation and terminal burr in the manufacturing process, this paper proposes a 3D visual inspection system based on a stereo vision with single camera. It is technically noted that the base line of this single camera stereo was precisely calibrated by the image processing procedure. Also to extract the measuring point coordinates for computing disparity; the error is reduced with original algorithm. Comparing its performance with that of human inspection using industrial microscope, the proposed 3D inspection could be an alternative in precision and in processing cost. Since the practical specification in 3D precision is less than 0.02 mm and the experimental performance was around the same, it was demonstrated by the proposed system that the soldering defect with terminal deformation and terminal burr in inspection, especially in 3D inspection, was decreased. In order to realize the inline inspection, this paper will suggest how the human inspection of the products could be modeled and be implemented by the computer system especially in manufacturing process.

This paper presents a method to recognize polygonal-shaped particles (i.e. rectangles, regular/irregular prisms) among agglomerated crystals from in situ images during a crystallization process. The aim is to measure the particle size distributions (PSD), which are key measurements needed for the purification operations and the quality control of chemical products or drugs. The method is first based on detecting the geometric features of the particles identified by their salient corners. A clustering technique is then applied by grouping three correspondent salient corners belonging to the same polygon. The efficiency of the proposed method is tested on particles of Ammonium Oxalate during batch crystallization in pure water. Particle size distributions are calculated, and a quantitative comparison between automatic and manual sizing is performed.

The progress in image acquisition technique provides life sciences with an abundance of data. Image analysis facilitates the assessment. The actin cytoskeleton plays a crucial role in understanding the behavior of osteoblastic cells on biomaterials. It can be visualized by confocal laser scanning microscopy. In these images it appears as a dense network of bright ridges which is so far only qualitatively assessed. For quantification there is a need for ridge detection techniques which provide a geometrical description of this graph-like feature. The state of the art methods do not cope with the systematical degradation of the brightness information. This paper presents the key part of a ridge tracking technique, which makes more efficient use of context information. Two random models confirm the accuracy and highlight the area for further improvement of the method.

This paper presents an algorithm enabling to detect disks on color images. The proposed method is based on a basic color segmentation giving a preliminary binary image. Then distance mapping is used to determine possible circles. And finally circle location is combined to color information in order to find the best-fitting disk. Furthermore, disks are, in a wide understanding, sets of points for which the distance to a particular point called center, is lower or equal to a given radius value. That is why the proposed method can also detect squares, octagons and other shapes that occur to be disks for a given distance, such as Euclidean, chessboard, Manhattan or chamfer distances. An application to pattern recognition for road sign interpretation is also presented in order to illustrate how road sign shape is a useful and significant information in the sign interpretation process.

Nano-level 3-D measurement is one of the key technologies for the current and future generation of production systems for semi-conductors, LCDs and nano-devices. To meet with these applications, wide range nano-level 3-D shape measurement method using combination of RGB laser lights has been developed. It measures the height of nano-objects from the combination of RGB LED lights interference. To analyze the combination of RGB lights, the color analysis method on xy-color plane has been introduced. In this method, the color changes on xy-color plane means the height changes. Experimental system to measure the three micro-meter height has been developed, and succeeded to measure the 50 nm step and 1000 nm step samples. The method has been applied to measure a nano-device, a contact needle for measurement. The shape of the needle has been extracted, successfully.

As a single-shot interferometric technique, spatial carrier interferometry has been thoroughly investigated, and it has been shown to have some problems, such as low spatial resolution. To overcome the problems, we propose a novel single-shot surface profiling technique that does not require carrier introduction. It is based on a model-fitting algorithm and estimates the model parameters and the heights of plural points simultaneously based on their multi-wavelength intensity data. The validity of the proposed method is demonstrated by computer simulations and actual experiments.

In this paper the problem of high performance software engineering is addressed in the context of image processing regarding productivity and optimized exploitation of hardware resources. Therefore, we introduce the functional array processing language Single Assignment C (SaC), which relies on a hardware virtualization concept for automated, parallel machine code generation. An illustrative benchmarking example proves both utility and adequacy of SaC for image processing.

In order to assess the quality of an X-ray imager it is necessary to measure the visibility of any artifact that might be present in the image. Several methods have been proposed in the literature to calculate this visibility. To predict the performance of these methods in the context of quality control of X-ray imagers, a base of 10 artifacts as different as possible in shape and aspect have been created (a pixel, a line, a step, various spots and noises). The amplitude for which each artifact has a probability of 50% to be detected has been determined. To do so, the artifacts have been observed merged with three noises of different spectra ("white noise", "high-frequency" noise and "low-frequency" noise). To determine the 50% detection probability amplitudes, a variant of the 2 Alternative Forced Choice procedure has been used. It has been checked that the measurement exploitation method is unbiased and its precision is sufficient. The dispersion of results between testers, around 15% in average, is also satisfactory. These results are a solid and objective basis to check the relevance and limits of visibility measurement methods described in literature, applied to the domain of quality control of X-ray imagers.

Defective unshu oranges (Citrus reticulate Blanco var. unshu) were sorted based on fluorescent imaging technique in a commercial packinghouse but fresh appearing unshu were rejected due to fluorescence appearing on their peel. We studied the various visible patterns based on colour, fluorescence and microscopic images, where even areas of the peel that are not obviously damaged can have fluorescence, to provide a categorization of fluorescence reasons. The categorization corresponded to: 1) hole and flow; 2) influenced by damaged or rotten fruits that have released peel oil onto it; 3) immature or poor peel quality; 4) whitish fluorescence due to agro-chemicals and 5) variation of the growing season. The identification of such patterns of fluorescence might be useful for citrus grading industry to take some initiatives to make the entire automated system more efficient.

Accurate quantitative reconstruction in kV cone-beam computed tomography (CBCT) is challenged by the presence of secondary radiations (scattering, fluorescence and bremsstrahlung photons) coming from the object and from the flat-panel detector itself. This paper presents a simulation study of the CBCT imaging chain as a first step towards the development of a comprehensive correction algorithm. A layer model of the detector is built in a Monte Carlo environment in order to help localizing and analyzing the secondary radiations. The contribution of these events to the final image is estimated with a forced-detection scheme to speed-up the Monte Carlo simulation without loss of accuracy. We more specifically assess to what extent a 2D description of the flat-panel detector would be sufficient for the forward model (i.e. the image formation process) of an iterative correction algorithm, both in terms of energy and incidence angle of incoming photons. A convolution model to account for detector secondary radiations is presented and validated. Results show that both object and detector secondary radiations have to be considered in CBCT.

The purpose of this study is to construct a model of an optical lens by using a multi camera array. It is known that virtually focused images can be produced by synthetic aperture focusing techniques. However there is a difference between the blur of the virtually focused image and the blur of an image produced by an optical lens. We suggest a method to correct this difference. Using our method, it is possible to create images that have multiple discrete focus depths, something that is impossible using an optical lens. Basic experiments were conducted, and the effectiveness of the approach was demonstrated.

The three dimensional X-ray computed tomography (3D-CT) has proved its successful usage as inspection method in non destructive testing. The generated 3D volume using high efficiency reconstruction algorithms contains all the inner structures of the inspected part. Segmentation of this volume reveals suspicious regions which need to be classified into defects or false alarms. This paper deals with the classification step using data fusion theory and support vector machine. Results achieved are very promising and prove the effectiveness of the data fusion theory as a method to build stronger classifier.

Many workpieces produced in large numbers and having a large variety of sizes and geometries, e.g. castings and forgings, have to be 100% inspected; on the one side, geometric tolerances need to be examined, and on the other side material defects, surface cracks have to be detected. In the paper a fully automated non-destructive testing technique is presented, whereby the workpiece is continuously moved and during this movement two measurements are carried out: first a thermographical measurement combined with inductive heating, where an infrared camera records the temperature distribution at the surface of the workpiece in order to localize material defects and surface cracks. In the second step a light sectioning measurement is carried out, to measure the 3d geometry of the piece. With help of registration the data from the two different sources are fused (merged) and evaluated together. The advantage of this technique is, that a more reliable decision can be made about the failures and their possible causes. The same registration technique can also be used for the comparison of different pieces and therefore to localize different failure types, compared to a 'golden', defect-free piece.

In human retina observation (with non mydriatic optical microscopes), a registration process is often employed to enlarge the field of view. For the ophthalmologist, this is a way to spare time browsing all the images. A lot of techniques have been proposed to perform this registration process, and indeed, its good evaluation is a question that can be raised. This article presents the use of the ε dissimilarity criterion to evaluate and compare some classical featurebased image registration techniques. The problem of retina images registration is employed as an example, but it could also be used in other applications. The images are first segmented and these segmentations are registered. The good quality of this registration is evaluated with the dissimilarity criterion for 25 pairs of images with a manual selection of control points. This study can be useful in order to choose the type of registration method and to evaluate the results of a new one.

We propose a mobile robot control method using 3D hand pose estimation without using sensors or controllers. The hand pose estimation we propose reduces the number of image features per data set, which makes the construction of a large-scale database possible, as well as estimation of the 3D hand poses of unspecified users with individual differences without sacrificing estimation accuracy. The system involves the construction in advance of a large database comprising three elements: hand joint information including the wrist, low-order proportional information on the hand images indicating the rough hand shape, and hand pose data comprised of low-order image features per data set.

This paper presents a new approach to non-rigid elastic registration. The method is applied to hyper spectral imaging data for the automatic quality control of decorative foils which are subject to deformation during lamination. A new image decimation procedure based on Savitzky-Golay smoothing is presented and applied in a multiresolution pyramid. Modified Fourier basis functions implemented by projection onto the orthogonal complement of a truncated Gram polynomial basis are presented. The modified functions are used to compute spectra whereby the Gibbs error associated with local gradients in the image are reduced. The paper also presents the first direct linear solution to weighted tensor product polynomial approximation. This method is used to regularize the patch coordinates, the solution is equivalent to a Galerkin type solution to a partial differential equations. The new solution is applied to published standard data set and to data acquired in a production environment. The speed of the new solution justifies explicit reference: the present solution implemented in MATLAB requires approximatly 1.3s to register an image of size 800 ×× 500 pixels. This is approximately a factor 10 to 100 faster than previously published results for the same data set.

In this paper we introduce a novel algorithm for automatic fault detection in textures. We study the problem of finding a defect in regularly textured images with an approach based on a template matching principle. We aim at registering patches of an input image in a defect-free reference sample according to some admissible transformations. This approach becomes feasible by introducing the so-called discrepancy norm as fitness function which shows particular behavior like a monotonicity and a Lipschitz property. The proposed approach relies only on few parameters which makes it an easily adaptable algorithm for industrial applications and, above all, it avoids complex tuning of configuration parameters. Experiments demonstrate the feasibility and the reliability of the proposed algorithms with textures from real-world applications in the context of quality inspection of woven textiles.

This paper deals with textured images and more particularly with directional textures. We propose a new parametric technique to estimate the orientation field of textures. It consists in partitioning the image into regions with homogeneous orientations, and then to estimate the orientation inside each of these regions, which allows us to maximize the size of the samples used to estimate the orientation without being corrupted by the presence of frontiers between regions. Once estimated the local - hence noisy - orientations of the texture using small filters (3×3 pixels), image partitioning is based on the minimization of the stochastic complexity (Minimum Description Length principle) of the orientation field. The orientation fluctuations are modeled with Von Mises probability density functions, leading to a fast and unsupervised partitioning algorithm. The accuracy of the orientations estimated with the proposed method is then compared with other approaches on synthetic images. An application to the processing of real images is finally addressed.

Current algorithms in automated indentation measurement in the context of Vickers microindentation hardness testing suffer from a lack of robustness with respect to entirely missed indentation corner points when applied to real world data sets. Four original algorithms are proposed, discussed and evaluated on a significant data set of indentation images. Three out of these four exhibit accuracy close to human operated hardness testing which has been conducted as a reference technique.

A large variety of autofocus functions used for assessing image sharpness is evaluated for the application in a passive autofocus system in the context of microindentation-based Vickers hardness testing. The functions are evaluated on a significant dataset of microindentation images with respect to the accuracy of sharpness assessment, their robustness to downsampling the image data, and their computational demand. Experiments suggest that the simple Brenner autofocus function is the best compromise between accuracy and computational effort in the considered application context.

In this paper, we present a novel approach for adaptive and progressive image acquisition, based on the progressive transmission of an image decomposed into compositions and superpositions of monovariate functions. The monovariate functions are iteratively constructed from the acquired data, to progressively reconstruct the final image: the transmission is performed directly in the 1D space of the monovariate functions, independently of any statistical properties of the image. Each monovariate function contains only a fraction of the pixels of the image. Each new transmitted monovariate function adds data to the previously transmitted monovariate functions. After each partial acquisition, by using the updated monovariate functions, the image is reconstructed with an increased resolution. Finally, once all the monovariate functions have been transmitted, the original image is reconstructed exactly at the maximum resolution of the sensor. This approach is characterized by its flexibility: any numbers of intermediate transmissions and reconstructions are possible. Moreover, the intermediate images can be reconstructed at any resolution, and for any number of intermediate reconstructions, the original image will be exactly reconstructed. Finally, the quantity of data is independent of the number and resolutions of intermediate reconstructions. Our contributions include the application of a flexible progressive transmission scheme to provide a progressive and flexible acquisition at various resolutions. Moreover, the accuracy of the full resolution image is preserved, and the acquired data are encrypted and resilient to packet-loss.

In this work, we intend to introduce a matching algorithm for 3D reconstruction of a planar object immersed in three dimensional scene with stereo pair of images. Such algorithm starts by matching closed contours having the same affine shape. This task is obtained by applying a complete and stable set of absolute affine invariants which have been computed after an affine invariant re-sampling procedure based on a normalized affine arc length. For each couple of corresponding contours, and by using the pseudo-inverse method, we introduce an efficient and robust affine motion estimation. The robustness is given by the contribution of all contour points to the geometrical parameters estimation to reconstruct apparent movement. We propose this idea to consolidate conventional matching algorithm in the context of stereovision assuming that objects are placed enough far from sensors. In this paper, we illustrate the efficiency of the proposed algorithm by applying it in archeological context.