Spectral Clustering Using Multilinear SVD: Analysis, Approximations and Applications

Authors: Debarghya Ghoshdastidar, Ambedkar Dukkipati

AAAI 2015 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental Experiments on geometric grouping and motion segmentation demonstrate the practical significance of the proposed methods. In this section, we conduct experiments on geometric grouping and motion segmentation.
Researcher Affiliation Academia Debarghya Ghoshdastidar and Ambedkar Dukkipati Department of Computer Science & Automation Indian Institute of Science Bangalore 560012, India email:{debarghya.g,ad}@csa.iisc.ernet.in
Pseudocode Yes Algorithm 1 Clustering using m-ary affinity relations, Algorithm 2 Column sampling variant of Algorithm 1, Algorithm 3 Nystr om approximation of Algorithm 1
Open Source Code No The paper does not provide any explicit statements about the availability of source code or links to a code repository.
Open Datasets Yes Finally, we conduct experiments on the Hopkins 155 motion segmentation database (Tron and Vidal 2007), where each video contains two or three independent motions.
Dataset Splits No The paper mentions using the 'Hopkins 155 motion segmentation database' but does not provide specific details on how this dataset was split into training, validation, or test sets.
Hardware Specification No The paper does not provide any specific details regarding the hardware (e.g., CPU, GPU models, memory) used to run the experiments.
Software Dependencies No The paper does not provide specific details on ancillary software, such as programming languages, libraries, or solvers with version numbers, used for implementation.
Experiment Setup Yes For geometric grouping... The m-way affinity is simply e cf for some parameter c > 0. For motion segmentation... we use 4th-order tensors, and fit group of four trajectories in a subspace of dimension 2. The affinities are of the form e cf( ), where f( ) is the fitting error. Algorithm 2... Number of sampled columns c; Threshold parameter τ > 0. Algorithm 3... Number of initial clusters kr; Number of vertices chosen from each cluster nr( (m 1)).