Graciela Lara

A computational model of perceptual saliency for 3D objects in virtual environments

When giving directions to the location of an object, people typically use other attractive objects as reference, that is, reference objects. With the aim to select proper reference objects, useful for locating a target object within a virtual environment (VE), a computational model to identify perceptual saliency is presented. Based on the object’s features with the major stimulus for the human visual system, three basic features of a 3D object (i.e., color, size, and shape) are individually evaluated and then combined to get a degree of saliency for each 3D object in a virtual scenario. An experiment was conducted to evaluate the extent to which the proposed measure of saliency matches with the people’s subjective perception of saliency; the results showed a good performance of this computational model.

A Computational Measure of Saliency of the Shape of 3D Objects

The shape of an object is a basic characteristic that when attracts the viewers’ attention represents a salient feature. In this paper we propose a computational measure of saliency of the shape of 3D objects in virtual reality, based on the proportion of empty and full space within its bounding box. This measure of saliency is part of a computational model aimed to the selection of appropriate reference objects to facilitate the location of objects within a 3D virtual environment. An experiment was conducted to understand to which extent the proposed measure of saliency matches with the people’s subjective perception of saliency; results showed a good performance of the metric.

3D objects’ shape relevance for saliency measure

The shape of an object is one of the features that attracts viewers’ attention, making it salient from other objects in a scenario. However, the shape of an object might be linked to the viewer’s personal experiences. In order to verify if the shape of an object is a feature not linked to the meaning that the object has for the viewer, in this paper we replicate an empirical study, using abstract object. 23 male and 17 female ordered objects according to the most attractive shapes. Using a computational measurement of saliency of the shape of 3D objects in Virtual Reality, based on the proportion of empty and full space within its bounding box, we found that our metric matches, when abstract or non-abstract objects are evaluated.

Computerized spatial language generation for object location

Spatial language is the syntax used for object or place locations. Because an object location is inherently relative, it implies a frame of reference, which in turn may be aided by a reference object, other than the one to be located. This reference object is commonly selected based on its perceptual salience, that is, its more prominent features. Computer systems linked to various research areas have been developed to facilitate the communication and/or interpretation of spatial language for localization tasks. In this paper is presented a literature review of computer systems that adopt spatial language and perceptual salience for object location.

Coverage of the university curricula for the Software Engineering industry in Mexico

Making to match the software industry requirements with the academy is a significant challenge that must be accomplished for the benefit of both sectors. In Mexico, the small and medium size companies (SMEs) are key for the software development, in such a way that having highly capable personal in the development of high quality software products is a fundamental piece to warranty their permanency in the market. In this context, the Mexican organizations for the accreditation to ensure quality in Software Engineering are: for the SMEs software industry, the Moprosoft norm; and for the undergraduate education institutions in Computer Science and Informatics, the CONAIC. In this document a comparative analysis is made to understand the coverage of the academic curricula in higher education, regarding the software industry requirements; based on what these government norms establish for their accreditation.

A Computational Measure of Saliency of the Texture Based a Saliency Map by Color

Understanding what attracts the human eye has always interested scientists; this represents the saliency concept. In this context, texture is a visual feature of the external surfaces of objects that can influence the senses, creating a force of attraction or rejection. When the texture of an object attracts the viewers’ attention it represents a salient feature. However, measuring how salient is an object texture is not trivial. In this paper we propose a computational measure of saliency of the visual texture, using a salience map by color. This measure of saliency is part of a computational model aimed to the selection of an appropriate reference objects to facilitate the location of objects within a 3D virtual environment. In order to verify this computational measurement, an experiment was conducted with 40 people, 30 male and 10 female, to understand to which extent the proposed measure of saliency matches with the people’s subjective perception of saliency. Results show a good metric performance.

Mexican Spanish Affective Dictionary

In the study of Affective Computing, the lexicon-based approach represents a useful mechanism that consists on using rated words to understand their affective role in a conversation. One of the most used lists of affectively rated words is the Affective Norms for English Words (ANEW), which evaluates the dimensions of pleasure, arousal and dominance for the English language. This list has been translated for other languages such as German or Spanish with effective results; however, there is not an affective lexicon for Mexican Spanish, rated by Mexicans. Based on the ANEW methodology, but using the most frequently words in Mexican Spanish language, similar to emoticons figures for the evaluation and an ad hoc app to collect the data, a list with means and standard deviation for Mexican Spanish words was obtained. Results and main differences with the ANEW study are here discussed.

Automatic Directions for Object Localization in Virtual Environments.

In order to assist users in the process of locating objects in Virtual Environments (VE), we automatize the process of giving directions through a computational model. This model generates directions in natural language by using spatial and perceptual aspects. It involves three main processes: (1) a computational model of perceptual saliency for 3D objects; (2) a user model and an explicit representation of virtual world semantics; and (3) the algorithm for automatic generation of directions to locate objects in natural language. Reference frames and reference objects support the model. For the selection of the best reference 3D object are considered three criteria: the perceptual saliency of the objects, the probability of the user to remember the object location, and prior knowledge from the user about the object. This paper presents the structure and the processes of the proposed model.

Framework for the Analysis of Smart Cities Models

Smart cities evolution forces auto adjustments. A constant change that difficult methodologies and tools development aimed to measure and evaluate the huge number of variables involved. The Smart City metrics model is composed by its determined key performed indicators (KPI); with different aims a number of models have been proposed by different organizations, which difficult its comparison. In this paper, we propose a framework to apply Data Science to KPIs from Open Data. This framework is organized by a set of tools: a KPI tree structure; a JSON document; a web app with non-supervised or supervised knowledge for the models evaluation; and the infrastructure for reports reception and attention. In such a way that this framework creates an infrastructure that goes from the treatment of Open Data to models evaluation and its management.

Scheme for the automatic generation of directions to locate objects in virtual environments

The automatic generation of direction in natural language, for the location of objects, is an ongoing research area heavily supported by the use of virtual environments (VEs). Important components of spatial language such as the selected reference object, along with specific features related to the situation of the scenario and the user, have to be properly combined in order to create a helpful direction to locate an object within a VE. In this paper we present a scheme, constructed upon literature review and specific empirical data, to link those different elements related to the location of objects, aimed to establish the suitable algorithms for the automatic generation of spatial language in VEs.