Robust Principal Component Analysis with Complex Noise

Authors: Qian Zhao, Deyu Meng, Zongben Xu, Wangmeng Zuo, Lei Zhang

ICML 2014 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental We evaluate the performance of the proposed Mo G-RPCA method on synthetic, face and video data. Two criteria were utilized for performance assessment. (1) Relative reconstruction error (RRE): ˆL L F / L F , where L and ˆL denote the ground truth and reconstructed low-rank matrices, respectively. (2) Estimated rank (ER): the rank of ˆL6. The performance of each competing method on each simulation was evaluated as the average over the 20 matrices in terms of RRE and ER, as listed in Table 1.
Researcher Affiliation Academia School of Mathematics and Statistics, Xi an Jiaotong University, Xi an, China School of Computer Science and Technology, Harbin Institute of Technology, Harbin, China Department of Computing, The Hong Kong Polytechnic University, Hong Kong, China
Pseudocode No No pseudocode or algorithm block found.
Open Source Code No No statement or link providing access to the authors' open-source code for Mo G-RPCA.
Open Datasets Yes The second subset of the Extended Yale B database (Georghiades et al., 2001), consisting of 64 faces of one subject with size 192 168, was used to generate the data matrix with size 32256 64. Four commonly utilized video sequences, including two indoor scenes (Bootstrap and Hall) and two outdoor scenes (Fountain and Campus), provided by Li et al. (2004)7, were adopted in our experiments.
Dataset Splits No The paper discusses data extraction but does not explicitly provide training/validation/test dataset splits or cross-validation setup.
Hardware Specification Yes All experiments were implemented in Matlab on a PC with 2.60GHz CPU and 8GB RAM.
Software Dependencies No The paper mentions 'Matlab' but does not provide specific version numbers for Matlab or any other software dependencies.
Experiment Setup Yes Throughout our experiments, we set µ0 = 0, and α01, . . . , α0K, β0, a0, b0, c0, d0 a small value 10 6.