SCOUT: Prefetching for Latent Structure Following Queries

Today's scientists are quickly moving from in vitro to in silico experimentation: they no longer analyze natural phenomena in a petri dish, but instead they build models and simulate them. Managing and analyzing the massive amounts of data involved in simulations is a major task. Yet, they lack the tools to efficiently work with data of this size. One problem many scientists share is the analysis of the massive spatial models they build. For several types of analysis they need to interactively follow the structures in the spatial model, e.g., the arterial tree, neuron fibers, etc., and issue range queries along the way. Each query takes long to execute, and the total time for executing a sequence of queries significantly delays data analysis. Prefetching the spatial data reduces the response time considerably, but known approaches do not prefetch with high accuracy. We develop SCOUT, a structure-aware method for prefetching data along interactive spatial query sequences. SCOUT uses an approximate graph model of the structures involved in past queries and attempts to identify what particular structure the user follows. Our experiments with neuroscience data show that SCOUT prefetches with an accuracy from 71% to 92%, which translates to a speedup of 4x-15x. SCOUT also improves the prefetching accuracy on datasets from other scientific domains, such as medicine and biology.

Analyse et exploitation des données de LIDAR aéroportés pour la caractérisation des milieux boisés de la Suisse

Gilles Gachet

Forest ecosystems show diverse and evolving structures. Accurate descriptions and a good knowledge of these environments are therefore a real challenge. Existing forest cover investigation methods are mostly laborious, hard to replicate and usually based on the skills of a small number of experts. Moreover, these methods are bound to standard two-dimensional information originated from aerial images or land surveys. The use of LiDAR elevation models leads to the development of new methodologies likely to propose innovative solutions to improve and ease the description and study of forests. In practice, managers often carry out maintenance forestry tasks as well as experimental ones. These tasks are successfully completed by means of tree observation and census. It is precisely with census that LiDAR technology can provide innovative solutions, allowing collection of accurate and reliable information from objects standing out of ground surface. Spatial analysis tools are then used to establish suitable indicators or cartographic documents to identify and define forest management objectives, to plan measures to achieve these goals, or even to monitor the results. The main objective of this PhD thesis is to propose new methods based on LiDAR technology applied to forestry. At first the study focuses on basic characteristics of a forest cover. Specific investigation of the potential of LiDAR elevation models to estimate the location, the height of individual trees and stands is carried out. In a second stage issues closer to management tasks are addressed, in particular analysis of horizontal and vertical structure of the forest, quantification of stock and identification of significant changes between two successive planting situations on the same site. Concerns of this study are to divert synthetic and normalized indicators in order to facilitate the actions and decision making process of the forester. It wants to extract innovative information in order to improve the knowledge and the description of wooden environments. A more or less significant participation of the skilled practitioner is required to configure the operations and to interpret the indicators, purpose of which is to save time by allowing more targeted and more complete ground investigation, subjective factor being in the benefit of the results. Although discussed issues as well as presented developments are based on common forest problems and of universal nature, they are presented in a context strongly influenced by Swiss forestry. The new contributions as well as the boundaries of the LiDAR technology are central to the reflection.

EPFL2009

Analyse et exploitation des données de LIDAR aéroportés pour la caractérisation des milieux boisés de la Suisse

Gilles Gachet

EPFL2009