Leandro A. F. Fernandes

Real-time line detection through an improved Hough transform voting scheme

The Hough transform (HT) is a popular tool for line detection due to its robustness to noise and missing data. However, the computational cost associated to its voting scheme has prevented software implementations to achieve real-time performance, except for very small images. Many dedicated hardware designs have been proposed, but such architectures restrict the image sizes they can handle. We present an improved voting scheme for the HT that allows a software implementation to achieve real-time performance even on relatively large images. Our approach operates on clusters of approximately collinear pixels. For each cluster, votes are cast using an oriented elliptical-Gaussian kernel that models the uncertainty associated with the best-fitting line with respect to the corresponding cluster. The proposed approach not only significantly improves the performance of the voting scheme, but also produces a much cleaner voting map and makes the transform more robust to the detection of spurious lines.

An Efficient Naturalness-Preserving Image-Recoloring Method for Dichromats

We present an efficient and automatic image-recoloring technique for dichromats that highlights important visual details that would otherwise be unnoticed by these individuals. While previous techniques approach this problem by potentially changing all colors of the original image, causing their results to look unnatural to color vision deficients, our approach preserves, as much as possible, the images original colors. Our approach is about three orders of magnitude faster than previous ones. The results of a paired-comparison evaluation carried out with fourteen color-vision deficients (CVDs) indicated the preference of our technique over the state-of-the-art automatic recoloring technique for dichromats. When considering information visualization examples, the subjects tend to prefer our results over the original images. An extension of our technique that exaggerates color contrast tends to be preferred when CVDs compared pairs of scientific visualization images. These results provide valuable information for guiding the design of visualizations for color-vision deficients.

A Physiologically-based Model for Simulation of Color Vision Deficiency

Color vision deficiency (CVD) affects approximately 200 million people worldwide, compromising the ability of these individuals to effectively perform color and visualization-related tasks. This has a significant impact on their private and professional lives. We present a physiologically-based model for simulating color vision. Our model is based on the stage theory of human color vision and is derived from data reported in electrophysiological studies. It is the first model to consistently handle normal color vision, anomalous trichromacy, and dichromacy in a unified way. We have validated the proposed model through an experimental evaluation involving groups of color vision deficient individuals and normal color vision ones. Our model can provide insights and feedback on how to improve visualization experiences for individuals with CVD. It also provides a framework for testing hypotheses about some aspects of the retinal photoreceptors in color vision deficient individuals.

An improved contrast enhancing approach for color-to-grayscale mappings

Despite the widespread availability of color sensors for image capture, the printing of documents and books are still primarily done in black-and-white for economic reasons. In this case, the included illustrations and photographs are printed in grayscale, with the potential loss of important information encoded in the chrominance channels of these images. We present an efficient contrast enhancement algorithm for color-to-grayscale image conversion that uses both luminance and chrominance information. Our algorithm is about three orders of magnitude faster than previous optimization-based methods, while providing some guarantees on important image properties. More specifically, our approach preserves gray values present in the color image, ensures global consistency, and locally enforces luminance consistency. Our algorithm is completely automatic, scales well with the number of pixels in the image, and can be efficiently implemented on modern GPUs. We also introduce an error metric for evaluating the quality of color-to-grayscale transformations.

A general framework for subspace detection in unordered multidimensional data

The analysis of large volumes of unordered multidimensional data is a problem confronted by scientists and data analysts every day. Often, it involves searching for data alignments that emerge as well-defined structures or geometric patterns in datasets. For example, straight lines, circles, and ellipses represent meaningful structures in data collected from electron backscatter diffraction, particle accelerators, and clonogenic assays. Also, customers with similar behavior describe linear correlations in e-commerce databases. We describe a general approach for detecting data alignments in large unordered noisy multidimensional datasets. In contrast to classical techniques such as the Hough transforms, which are designed for detecting a specific type of alignment on a given type of input, our approach is independent of the geometric properties of the alignments to be detected, as well as independent of the type of input data. Thus, it allows concurrent detection of multiple kinds of data alignments, in datasets containing multiple types of data. Given its general nature, optimizations developed for our technique immediately benefit all its applications, regardless the type of input data.

Association Rule Visualization and Pruning Through Response-Style Data Organization and Clustering

Association rules are a very popular non-supervised data mining technique for extracting co-relation in large set of data transactions. Although the vast use, the analysis of mined rules may be intricate for non-experts, and the technique effectiveness is constrained by the data dimensionality. This paper presents a pre-processing approach that uses (1) dual scaling in order to present the mined rules with some semantic contextualization that assists interpretation, and (2) mean shift clustering to reduce data dimensionality. We tested our model with real data collected from accident reports in petroleum industry.

Geometric Algebra: A Powerful Tool for Solving Geometric Problems in Visual Computing

Geometric problems in visual computing (computer graphics, computer vision, and image processing) are typically modeled and solved using linear algebra (LA). Thus, vectors are used to represent directions and points in space, while matrices are used to model transformations. LA, however, presents some well-known limitations for performing geometric computations. As a result, one often needs to aggregate different formalisms (e.g., quaternions and Plücker coordinates) to obtain complete solutions. Unfortunately, such extensions are not fully compatible among themselves, and one has to get used to jumping back and forth between formalisms, filling in the gaps between them. Geometric algebra (GA), on the other hand, is a mathematical framework that naturally generalizes and integrates useful formalisms such as complex numbers, quaternions and Plücker coordinates into a high-level specification language for geometric operations. Due to its consistent structure, GA equations are often universal and generally applicable. They extend the same solution to higher dimensions and to all kinds of geometric elements, without having to handle special cases, as it happens in conventional techniques. This tutorial aims at introducing the fundamental concepts of GA as a powerful mathematical tool to describe and solve geometric problems in visual computing.

A fast and accurate approach for computing the dimensions of boxes from single perspective images

This paper describes an accurate method for computing the dimensions of boxes directly from perspective projection images acquired by conventional cameras. The approach is based on projective geometry and computes the box dimensions using data extracted from the box silhouette and from the projection of two parallel laser beams on one of the imaged faces of the box. In order to identify the box silhouette, we have developed a statistical model for homogeneous-background-color removal that works with a moving camera, and an efficient voting scheme for the Hough transform that allows the identification of almost collinear groups of pixels. We demonstrate the effectiveness of the proposed approach by automatically computing the dimensions of real boxes using a scanner prototype that implements the algorithms and methods describe din the paper. We also present a discussion of the performed measurements, and an error propagation analysis that allows the method to estimate, from each single video frame, the uncertainty associated to all measurements made over thatframe, in real-time.

Determining the location of buildings given a single picture, environment maps and inaccurate GPS coordinates

A recurring task for travelers and citizens of large cities is to obtain information about historical monuments, interesting façades, and business establishments. Even in front of the target building, a traveler may experience some difficulties in getting information about the place of interest due to, for instance, language issues. In addition, even with the help of a GPS-enabled smartphone, the identification of the establishment may be difficult due to the lack of accuracy of the positioning system, which may not provide precise location of the site. We present a computational process that is capable of identifying the correct geographic coordinates of the place of interest given a single photo image of a target building. We use the inaccurate GPS coordinates as start point for automated visual search into first-person web mapping services (e.g., Google Street View, Microsoft StreetSide). By going through the views of the streets, our system looks for the façade presented in the input image and retrieve its location. We demonstrate the proposed approach and algorithms by building a prototype of the application. Our solution employs graph traversal algorithms, computer vision and image processing techniques such as ASIFT, projection, triangulation, and histogram analysis. The effectiveness of our solution was asserted by automatically retrieving the geographic coordinates of façades in our hometown.

Corrigendum: Corrigendum to Real-time line detection through an improved hough transform voting scheme [Pattern Recognition 41 (1) 299-314]

Clarification A vertical line in the original imagewould cause a division by zero operation that is not dealtwithin Algorithm2, line 20 ( / m′ =xu/ √ 1− x2 v in Eq. (14)). In such a case, however, the / m′ is zero, and 2 and (Eq. (13)) are also equal to zero. Thus, in practice, division by zero never actually happens. Acknowledgments: The authors would like to thank Ian (from University of Waterloo) for asking this.