Maximilian Scherer

Sketch-based 3D model retrieval using diffusion tensor fields of suggestive contours

The number of available 3D models in various areas increase steadily. Effective methods to search for those 3D models by content, rather than textual annotations, are crucial. For this purpose, we propose a new approach for content based 3D model retrieval by hand-drawn sketch images. This approach to retrieve visually similar mesh models from a large database consists of three major steps: (1) suggestive contour renderings from different viewpoints to compare against the user drawn sketches; (2) descriptor computation by analyzing diffusion tensor fields of suggestive contour images or the query sketch respectively; (3) similarity measurement to retrieve the models and the most probable view-point from which a model was sketched. Our proposed sketch based 3D model retrieval system is very robust against variations of shape, pose or partial occlusion of the user draw sketches. Experimental results are presented and indicate the effectiveness of our approach for sketch-based 3D mode retrieval.

SHREC'12 track: sketch-based 3D shape retrieval

Sketch-based 3D shape retrieval has become an important research topic in content-based 3D object retrieval. The aim of this track is to measure and compare the performance of sketch-based 3D shape retrieval methods implemented by different participants over the world. The track is based on a new sketch-based 3D shape benchmark, which contains two types of sketch queries and two versions of target 3D models. In this track, 7 runs have been submitted by 5 groups and their retrieval accuracies were evaluated using 7 commonly used retrieval performance metrics. We hope that the benchmark, its corresponding evaluation code, and the comparative evaluation results of the state-of-the-art sketch-based 3D model retrieval algorithms will contribute to the progress of this research direction for the 3D model retrieval community.

STELA: sketch-based 3D model retrieval using a structure-based local approach

Since 3D models are becoming more popular, the need for effective methods capable of retrieving 3D models are becoming crucial. Current methods require an example 3D model as query. However, in many cases, such a query is not easy to get. An alternative is using a hand-draw sketch as query. We present a structure-based local approach (STELA) for retrieving 3D models using a rough sketch as query. It consists of four steps: get an abstract image, detect keyshapes, compute a local descriptor, and match local descriptors. We represent a 3D model by means of suggestive contours. Our proposal includes an additional step aiming at reducing the number of models that will be compared by our local approach. The proposed method is invariant to position, scale, and rotation changes as well. We evaluate our method using the first-tier precision and compare it with a current global approach (HELO). Our results show an increasing in precision for many classes of 3D models.

Sketch-based 3D model retrieval using keyshapes for global and local representation

Since 3D models are becoming more popular, the need for effective methods capable of retrieving 3D models is becoming crucial. Current methods require an example 3D model as query. However, in many cases, such a query is not easy to get. An alternative is using a hand-drawn sketch as query. In this work, we present a new keyshape based approach named HKO-KASD for retrieving 3D models using rough sketches as queries. Our approach comprises two general steps. First, a global descriptor is used to determine the appropriate viewpoint for each model. Second, we apply a local matching process to determine the final ranking for an input sketch. To this end, we present a local descriptor capable of working with sketch representations. The global descriptors as well as the local descriptors rely on a set of keyshapes precomputed from 2D representations of 3D models and from the query sketch as well. We evaluate our method using the first-tier precision and compare it with current approaches (HELO, STELA). Our results show a significant increase in precision for many classes of 3D models.

Content-based layouts for exploratory metadata search in scientific research data

Todays digital libraries (DLs) archive vast amounts of information in the form of text, videos, images, data measurements, etc. User access to DL content can rely on similarity between metadata elements, or similarity between the data itself (content-based similarity). We consider the problem of exploratory search in large DLs of time-oriented data. We propose a novel approach for overview-first exploration of data collections based on user-selected metadata properties. In a 2D layout representing entities of the selected property are laid out based on their similarity with respect to the underlying data content. The display is enhanced by compact summarizations of underlying data elements, and forms the basis for exploratory navigation of users in the data space. The approach is proposed as an interface for visual exploration, leading the user to discover interesting relationships between data items relying on content-based similarity between data items and their respective metadata labels. We apply the method on real data sets from the earth observation community, showing its applicability and usefulness.

Retrieval and exploratory search in multivariate research data repositories using regressional features

Increasing amounts of data are collected in many areas of research and application. The degree to which this data can be accessed, retrieved, and analyzed is decisive to obtain progress in fields such as scientific research or industrial production. We present a novel method supporting content-based retrieval and exploratory search in repositories of multivariate research data. In particular, functional dependencies are a key characteristic of data that researchers are often interested in. Our methods are able to describe the functional form of such dependencies, e.g., the relationship between inflation and unemployment in economics. Our basic idea is to use feature vectors based on the goodness-of-fit of a set of regression models, to describe the data mathematically. We denote this approach Regressional Features and use it for content-based search and, since our approach motivates an intuitive definition of interestingness, for exploring the most interesting data. We apply our method on considerable real-world research datasets, showing the usefulness of our approach for user-centered access to research data in a Digital Library system.

The PROBADO project: approach and lessons learned in building a digital library system for heterogeneous non-textual documents

The PROBADO project is a research effort to develop and operate advanced Digital Library support for non-textual documents. The main goal is to contribute to all parts of the Digital Library work flow from content acquisition over indexing to search and presentation. While not limited in terms of supported document types, reference support is developed for classical digital music and 3D architectural models. In this paper, we review the overall goals, approaches taken, and lessons learned so far in a highly integrated effort of university researchers and library experts. We address the problem of technology transfer, aspects of repository compilation, and the problem of inter-domain retrieval. The experiences are relevant for other project efforts in the nontextual Digital Library domain.

Guided Sketching for Visual Search and Exploration in Large Scatter Plot Spaces

Recently, there has been an interest in methods for filtering large scatter plot spaces for interesting patterns. However, user interaction remains crucial in starting an explorative analysis in a large scatter plot space. We introduce an approach for explorative search and navigation in large sets of scatter plot diagrams. By means of a sketch-based query interface, users can start the exploration process by providing a visual example of the pattern they are interested in. A shadow-drawing approach provides suggestions for possibly relevant patterns while query drawing takes place, supporting the visual search process. We apply the approach on a large real-world data set, demonstrating the principal functionality and usefulness of our technique.

Guided discovery of interesting relationships between time series clusters and metadata properties

Visual cluster analysis provides valuable tools that help analysts to understand large data sets in terms of representative clusters and relationships thereof. Often, the found clusters are to be understood in context of belonging categorical, numerical or textual metadata which are given for the data elements. While often not part of the clustering process, such metadata play an important role and need to be considered during the interactive cluster exploration process. Traditionally, linked-views allow to relate (or loosely speaking: correlate) clusters with metadata or other properties of the underlying cluster data. Manually inspecting the distribution of metadata for each cluster in a linked-view approach is tedious, especially for large data sets, where a large search problem arises. Fully interactive search for potentially useful or interesting cluster to metadata relationships may constitute a cumbersome and long process. To remedy this problem, we propose a novel approach for guiding users in discovering interesting relationships between clusters and associated metadata. Its goal is to guide the analyst through the potentially huge search space. We focus in our work on metadata of categorical type, which can be summarized for a cluster in form of a histogram. We start from a given visual cluster representation, and compute certain measures of interestingness defined on the distribution of metadata categories for the clusters. These measures are used to automatically score and rank the clusters for potential interestingness regarding the distribution of categorical metadata. Identified interesting relationships are highlighted in the visual cluster representation for easy inspection by the user. We present a system implementing an encompassing, yet extensible, set of interestingness scores for categorical metadata, which can also be extended to numerical metadata. Appropriate visual representations are provided for showing the visual correlations, as well as the calculated ranking scores. Focusing on clusters of time series data, we test our approach on a large real-world data set of time-oriented scientific research data, demonstrating how specific interesting views are automatically identified, supporting the analyst discovering interesting and visually understandable relationships.

A benchmark for content-based retrieval in bivariate data collections

Huge amounts of various research data are produced and made publicly available in digital libraries. An important category is bivariate data (measurements of one variable versus the other). Examples of bivariate data include observations of temperature and ozone levels (e.g., in environmental observation), domestic production and unemployment (e.g., in economics), or education and income level levels (in the social sciences). For accessing these data, content-based retrieval is an important query modality. It allows researchers to search for specific relationships among data variables (e.g., quadratic dependence of temperature on altitude). However, such retrieval is to date a challenge, as it is not clear which similarity measures to apply. Various approaches have been proposed, yet no benchmarks to compare their retrieval effectiveness have been defined. In this paper, we construct a benchmark for retrieval of bivariate data. It is based on a large collection of bivariate research data. To define similarity classes, we use category information that was annotated by domain experts. The resulting similarity classes are used to compare several recently proposed content-based retrieval approaches for bivariate data, by means of precision and recall. This study is the first to present an encompassing benchmark data set and compare the performance of respective techniques. We also identify potential research directions based on the results obtained for bivariate data. The benchmark and implementations of similarity functions are made available, to foster research in this emerging area of content-based retrieval.