Penalized preimage learning in kernel principal component analysis

Keyphrases

• Kernel Principal Component Analysis 100%
• Feature Vector 50%
• Face Image 50%
• Penalization 50%
• Learning Process 50%
• Combination Weight 50%
• Convexity Constraint 50%
• Optimization Function 50%
• Facial Expression 25%
• Occlusion 25%
• Image Processing 25%
• Laplacian 25%
• Image Denoising 25%
• Visual Quality 25%
• Image Data 25%
• Well-defined 25%
• Learning Algorithm 25%
• Existing Techniques 25%
• Unsupervised Learning 25%
• Mean Square Error 25%
• Illumination Normalization 25%
• Kernel Methods 25%
• Penalization Method 25%
• Kernel Feature Space 25%
• Weighted Combination 25%
• Missing Parts 25%
• Expression Normalization 25%
• Penalized Function 25%
• Ridge Penalty 25%
• How-to Guide 25%
• Reconstructed Image Quality 25%
• Conditional mutual Information 25%

Computer Science

• Component Analysis 100%
• Principal Component 100%
• Learning Process 50%
• Feature Vector 50%
• Experimental Result 25%
• Facial Expression 25%
• Laplace Operator 25%
• image denoising 25%
• Learning Algorithm 25%
• Visual Quality 25%
• Reconstructed Image 25%
• Kernel Method 25%
• Mutual Information 25%
• Kernel Feature Space 25%

Mathematics

• Principal Component Analysis 100%
• Pre-Image 100%
• Feature Vector 13%
• Function Optimization 13%
• Conditionals 6%
• Laplace Operator 6%
• Image Data 6%
• Approximates 6%
• Input Data 6%
• Mean Square Error 6%
• Kernel Method 6%
• Data Space 6%
• Open Problem 6%
• Pointwise 6%
• Ridge 6%
• Mutual Information 6%
• Feature Space 6%
• Observed Sample 6%
• Reconstructed Image 6%