Uncorrelated Multi-View Discrimination Dictionary Learning for Recognition

Authors: Xiao-Yuan Jing, Rui-Min Hu, Fei Wu, Xi-Lin Chen, Qian Liu, Yong-Fang Yao

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

Reproducibility Variable Result LLM Response
Research Type Experimental Experiments on several public datasets demonstrate the effectiveness of the proposed approach.
Researcher Affiliation Academia 1State Key Laboratory of Software Engineering, School of Computer, Wuhan University, China 2National Engineering Research Center for Multimedia Software, School of Computer, Wuhan University, China 3College of Automation, Nanjing University of Posts and Telecommunications, China 4Institute of Computing Technology, Chinese Academy of Sciences, China
Pseudocode Yes Algorithm 1. UMD2L
Open Source Code No The paper does not provide concrete access to source code for the methodology described.
Open Datasets Yes The proposed UMD2L approach is verified on the Multi PIE (Cai et al., 2006), AR (Martinez and Benavente 1998), COIL-20 (Murase and Nayar 1995) and MNIST (Le Cun et al., 1998) datasets.
Dataset Splits Yes In all experiments, the tuning parameters in UMD2L ( O in dictionary learning phase, and J in classification phase) and the parameters of all compared methods are evaluated by 5-fold cross validation to avoid over-fitting.
Hardware Specification No The paper does not provide specific hardware details (e.g., exact GPU/CPU models, memory amounts) used for running its experiments.
Software Dependencies No The paper mentions techniques like PCA transformation, Gabor transformation, Karhunen-Loeve (KL) transformation, and Local Binary Patterns (LBP), but does not specify software names with version numbers.
Experiment Setup Yes Concretely, the parameters of UMD2L are set as 0.005 O and 0.001 J for three datasets. In addition, the default dictionary atoms number for each view in UMD2L is set as the number of training samples.