Unsupervised Feature Learning for 3D Scene Reconstruction with Occupancy Maps

Authors: Vitor Guizilini, Fabio Ramos

AAAI 2017 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental Experiments conducted in simulated and real large-scale datasets show a substantial gain in performance, while decreasing the amount of stored information by orders of magnitude without sacrificing accuracy.
Researcher Affiliation Academia Vitor Guizilini, Fabio Ramos School of Information Technologies, University of Sydney {vitor.guizilini,fabio.ramos}@sydney.edu.au
Pseudocode Yes Algorithm 1 ASK-Means initialization algorithm
Open Source Code Yes 1A C++ demo of the proposed algorithm is available in https: //bitbucket.org/vguizilini/cvpp
Open Datasets Yes In this section we validate the proposed algorithm using three different datasets: Room, a simulated indoor environment; Corridor, a real indoor environment; and Outdoor, a real outdoor environment2. ... 2Both real datasets were obtained from http://kos.informatik. uni-osnabrueck.de/3Dscans/
Dataset Splits No The paper mentions 'different ratios of training/testing points' and 'every training/testing ratio' but does not provide specific percentages or counts for a fixed training/validation/test split.
Hardware Specification No The paper does not provide any specific hardware details such as GPU or CPU models used for running the experiments.
Software Dependencies No The paper mentions various algorithms and methods (e.g., K-means, Hilbert Maps, kd-tree) and cites their original papers, but it does not specify any software dependencies with version numbers (e.g., programming languages, libraries, or frameworks with their versions) used for implementation or experimentation.
Experiment Setup Yes In all experiments, ASK-means (Sec. ) was used with a threshold of 0.1% the maximum distance between points. ... Unoccupied points are generated by randomly sampling the beams that produced occupied points (a ratio of 1 point / 2 meters was used throughout the paper).