Scalable Exploration of Highly Detailed and Annotated 3D Models

IEEE Transactions on Visualization and Computer Graphics, 2015

With broader availability of large-scale 3D model repositories, the need for efficient and effective exploration becomes more and more urgent. Existing model retrieval techniques do not scale well with the size of the database since often a large number of very similar objects are returned for a query, and the possibilities to refine the search are quite limited. We propose an interactive approach where the user feeds an active learning procedure by labeling either entire models or parts of them as "like" or "dislike" such that the system can automatically update an active set of recommended models. To provide an intuitive user interface, candidate models are presented based on their estimated relevance for the current query. From the methodological point of view, our main contribution is to exploit not only the similarity between a query and the database models but also the similarities among the database models themselves. We achieve this by an offline pre-processing stage, where global and local shape descriptors are computed for each model and a sparse distance metric is derived that can be evaluated efficiently even for very large databases. We demonstrate the effectiveness of our method by interactively exploring a repository containing over 100K models.

This paper presents a set of tools developed to model, visualize and analyze large 3D datasets built from 2D and 3D sensor data. These tools, grouped under the Atelier3D framework, first include a technique for processing and interactively visualizing datasets made of hundreds of millions 3D samples and tens of gigabytes of texture from digital photography. It also includes various analysis tools to visualize and transform the data, including non-photorealistic rendering techniques implemented on the GPU that provide the capabilities to extract information from the datasets not always visible using classical shading techniques. Atelier3D also contains advanced display and interaction functionalities to support multi-projector and multi-resolution tele-collaborative hardware configurations.

2009

A joint project between the Interdept. Research Center of the University of Padova (CIRGEO) andDr. Richard. M. Levy, teaching Professor at the Faculty of Environmental D esign of the University of Calgary (Canada), has been undertaken aimed to the generation ofa full D model of an historical building and the surrounding environment, based on a terrestrial laser scanning survey. The main,goal of this project is to provid e a D representation where,two different contents are merged,together: the object ’s geometry on one hand and a set of related historical and cultural information on the other hand. Indeed, through suitable VR authoring tools, like the Virtools Dev. software, it is now possible to build VR environments around laser scanning-based 3D models, which allow the user with a certain level of interactionwith the model itself. This solution opens interesting perspectives towards the use of 3D models as a mean to pr omote the national Cultural Heritage content among remote peo...

Lecture Notes in Computer Science, 2009

Most information visualisation methods are based on abstract visual representations without any concrete manifestation in the "real world". However, a variety of abstract datasets can indeed be related to, and hence enriched by, real-world aspects. In these cases an additional virtual representation of the 3D object can help to gain a better insight into the connection between abstract and real-world issues. We demonstrate this approach with two prototype systems that combine information visualisation with 3D models in multiple coordinated views. The first prototype involves the visualisation of in-car communication traces. The 3D model of the car serves as one view among several and provides the user with information about the car's activities. LibViz, our second prototype, is based on a full screen 3D representation of a library building. Measured data is visualised in overlaid, semi-transparent windows to allow the user interpretation of the data in its spatial context of the library's 3D model. Based on the two prototypes, we identify the benefits and drawbacks of the approach, investigate aspects of coordination between the 3D model and the abstract visualisations, and discuss principals for a general approach.

2016

Authors listed alphabetically We present ShapeNet: a richly-annotated, large-scale repository of shapes represented by 3D CAD models of ob-jects. ShapeNet contains 3D models from a multitude of semantic categories and organizes them under the Word-Net taxonomy. It is a collection of datasets providing many semantic annotations for each 3D model such as consis-tent rigid alignments, parts and bilateral symmetry planes,

eprints.utm.my

Currently, variety of software is already capable of handling a wide range of spatial problems, beginning with approaches for describing spatial objects to quite complex analysis and 3D visualisation. However, increasing number of applications need more advanced ...

1997

Every computer-rendered image requires three essential components: a 3D scene description, one or more sources of light, and a description of the camera or eye viewing the scene. The scene description is typically composed of one or more models, or 3D structures. Typically we think of a model as a stand-alone part, e.g. a pencil or a tree, and the scene as the assembly of these parts into a complete 3D environment. This attitude re ects the most common procedure for building up a 3D scene: one builds many models, and then assembles them. Each model contains two descriptions: a mathematical representation of the structure of the shape, and a recipe for guring out how that shape would look if illuminated. Let us examine these descriptions in turn.

The ShapeAnnotator is a graphical tool designed to assist an expert user in the task of annotating a surface mesh with concepts belonging to a domain of expertise. The user is first required to identify meaningful surface features; Each such feature can then be "labeled" with a concept, and the surface mesh, along with the features identified and their labels, can be saved.

Studies in Digital Heritage

The knowledge and study of built heritage is now deeply connected to methodologies associated with the capture of surface details via the production of point-data. These methodologies enable researchers to gather a wider range of information, which is increasingly more connected to technological advances. Such approaches influence the management of data, and these data are often redundant due to the ways in which they are captured. Massive data capture does not include preliminary selection based on metric, geometric, and material features of the object. A multi-scalar approach, in which the criteria for data capture depends on the goals of the survey, is needed to optimize the relationship between information and the scale of the models to be built. This case study involving a selection of fountains in Rome aims to apply these principles to urban contexts defined by a strong spatial connection between architectural and sculptural elements. Survey can express this distinctiveness th...