Alexandre Gonçalves Silva

Efficient computation of new extinction values from extended component tree

A gray-scale image can be interpreted as a topographical surface, and represented by a component tree, based on the inclusion relation of connected components obtained by threshold decomposition. Relations between plateaus, valleys or mountains of this relief are useful in computer vision systems. An important definition to characterize the topographical surface is the dynamics, introduced by Grimaud (1992), associated with each regional minimum. This concept has been extended, by Vachier and Meyer (1995), by the definition of extinction values associated with each extremum of the image. This paper proposes three new extinction values - two based on the topology of the component tree: (i) number of descendants and (ii) sub-tree height; and one geometric: (iii) level component bounding box (subdivided into extinctions of height, width or diagonal). This paper describes an efficient computation of these extinction values based on the incremental determination of attributes from the component tree construction in quasi-linear time, compares the computation time of the method and illustrates the usefulness of these new extinction values from real examples.

Classification of microstructures by morphological analysis and estimation of the hydration degree of cement paste in concrete

Cement is an essential material for the development of the civil construction. The analysis of cement-based materials is useful in the evaluation of mechanical properties, water absorption, air permeability and evolution of hydration degree. The latter is the motivation of this paper. New advantageous techniques of analysis based on the image information are possible using electronic microscopy. We propose here an alternative procedure to the segmentation of microstructures of cement and aggregates in concrete. The central idea consists of a particular use of the morphological operator watershed. We report the results of the analysis of about 450 scanning electron microscope (SEM) images.

Enriched Barcodes applied in Mobile Robotics and Augmented Reality

Bar codes are applied in several tasks to the automation of data acquisition through electronic devices. Withal, markers are used by applications of Augmented Reality and Mobile Robotics. This paper is a survey of applications made in both areas and proposes the development of a marker that incorporates the features of a barcode in two dimensions. Thus, the development process of the enriched bar code is displayed in details, and are proposed applications in augmented reality and mobile robotics. In augmented reality are made two applications to validate acquisition of position, orientation and data, and in the mobile robotics is made an application developed from conception, to design, control strategy, implementation and testing. Finally, advantages of the proposed technology and their limitations are presented.

Maps of distances for evaluating the tracking and measurement of electrical arcs by 3D-snake

This paper presents a 3D reconstruction (by 3D-snake) of the longitudinal axis of an electrical arc (discharge) in evolution based on stereo pairs of images. The accuracy of such tracking depends strongly on the halting of the iterative relaxation process that adjusts the 3D-snake. The key aspect to make such decision is detecting when each projection of the 3D-snake matches special features in the pair of images. For this sake, this paper proposes a simple and effective strategy based on the projection matrices of the cameras and tables named Maps of Distances (MDs) which are obtained by image processing.

Segmentation of Foveal Avascular Zone of the Retina Based on Morphological Alternating Sequential Filtering

The non-invasive visualization of retinal microvasculature is a traditional way to diagnose and predict some diseases. The region of foveal avascular zone (FAZ) is an important tool to quantify the macular ischemic with implications in the visual acuity. In this work, we proposed an automatic segmentation of FAZ using high-resolution retinal images captured by a retinography conjugated to Retinal Function Imager (RFI) apparatus. The pixels corresponding to FAZ are automatically detected, segmented and measured (area, perimeter, height and width). The proposed algorithm was tested with 20 images of 10 healthy volunteers, and the results were compared to three manual segmentations. The difference between the software and the manual segmentation was 12.8930% (the mean difference between humans was 7.8288%), and the time of each image segmentation was performed 14.1549 times faster by the computer. The method obtained accuracy of 0.9947, sensitivity of 0.8442, and specificity of 0.9972.

3D Reconstruction from Planar Contours: Analysis of Heuristic Tiling Approaches

3D reconstruction from planar contours using the heuristic approach implements rules for three different phases of the process: correspondence, tiling and branching. We have analyzed existing algorithms and have been able to isolate their constituent pieces which allowed us to foresee many other possible (atomic) contributions. These pieces of algorithms were implemented independently and different mixes were compared using performance, geometrical and user-centered metrics. It was found that user-centered analysis are not reliable; that local criterion do not reflect on the whole model’s quality and; that there is a tradeoff between performance and geometrical metrics. We could also find a particular mix of algorithm pieces that lead to a novel 3D reconstruction algorithm. Moreover, we have built an open source freeware framework where more mixes can be composed and where further testing and improvements could be carried out.

Adessowiki: Collaborative Scientific Programming Environment

Adessowiki is a collaborative environment for teaching and research in image processing. Adessowiki is composed of a collection of collaborative web pages in the form of a wiki. The articles of this wiki can embed programming code that will be executed on the server when the page is rendered, incorporating the results as figures, texts and tables on the document. The integrated collaborative environment of Adessowiki, containing documentation, programming code and execution results is able to create several possibilities of applications. This paper presents some of the applications where Adessowiki has been used, such as Scientific Writing and Virtual Learning Environment.

k-max: segmentation based on selection of Max-tree deep nodes

This work proposes the segmentation of grayscale image from of its hierarchical region based representation. The Maxtree structure has demonstrated to be useful for this purpose, offering a semantic vision of the image, therefore, reducing the number of elements to process in relation to the pixel based representation. In this way, a particular searching in this tree can be used to determine regions of interest with lesser computational effort. A generic application of detection of peaks is proposed through searching nodes to kup steps from leaves in the Max-tree (this operator will be called k-max), being each node corresponds to a connected component. The results are compared with the optimal thresholding and the H-maxima technique.

ratoca - Mouse by Detection of Markers Using a Camera of Video in Not Controlled Environment

This work approaches an alternative man-machine interface using a low cost camera. A mouse architecture by webcam is elaborated from the acquisition of images and processing of simplified markers defined appropriately. An useful equipment is intended in the most varied kinds of environments, with significant tolerance of illumination.

Detection of lines using hierarchical region based representation

This paper suggests an hierarchical region based representation to implement detection of shapes in gray-scale images. The max-tree structure has been demonstrated to be useful for this purpose, offering a semantic objects vision of the image, thereby reducing the number of elements to process in relation to the pixel based representation. In this way, particular thresholding information, associated with a specific shape analysis, can be used to determine the segmentation or filtering of regions of interest with less computational effort. To demonstrate this technique, a generic application of detection of lines is proposed using the distance transform approach on each connected component from the depth searching in the mentioned tree.