Multi-View Point Registration via Alternating Optimization

Authors: Junchi Yan, Jun Wang, Hongyuan Zha, Xiaokang Yang, Stephen Chu

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

Reproducibility Variable Result LLM Response
Research Type Experimental Extensive empirical evaluations of peer methods on both synthetic data and real images suggest our method is robust to large disturbance.
Researcher Affiliation Collaboration 1Department of Electronic Engineering, Shanghai Jiao Tong University, Shanghai, 200240, China 2IBM Research China, Shanghai, 201203, China 3Software Engineering Institute, East China Normal University, Shanghai, 200062, China 4College of Computing, Georgia Institute of Technology, Atlanta, Georgia, 30332, USA 5Institue of Data Science and Technology, Alibaba Group, Seattle, WA, 98101, USA
Pseudocode Yes Algorithm 1 Alternating concave optimization for three-view point registration
Open Source Code No No explicit statement or link indicating the release of source code for the methodology was found.
Open Datasets Yes The synthetic data used in this paper is generated by the template from Chui-Rangarajan data sets (Chui and Rangarajan 2003). The tested real image data is from CMU hotel and house sequences. The other two sequences (volvo and house) from the pose database (Vikstn et al. 2009).
Dataset Splits No The paper mentions using synthetic data generated from Chui-Rangarajan data sets and real image data from CMU hotel and house sequences and the pose database, but does not specify the train/validation/test splits or percentages used.
Hardware Specification Yes We implement all the competing methods in Matlab R2009 on a desktop with dual 2.53GHz CPU and 3G memory.
Software Dependencies Yes We implement all the competing methods in Matlab R2009 on a desktop with dual 2.53GHz CPU and 3G memory.
Experiment Setup No The paper states using '2-D similarity transformation' for rotation-invariant tests and 'affine transformation' for rotation-sensitive tests, but does not provide specific hyperparameter values or detailed system-level training settings.