A novel semi-supervised methodology for extracting tumor type-specific MRS sources in human brain data

Keyphrases

• Tumor Subtype 100%
• Human Brain 100%
• Type-specific 100%
• Brain Data 100%
• Single Voxel 66%
• Matrix Factorization 66%
• Class Information 66%
• Brain Tumor 33%
• Model Optimization 33%
• Tumor 33%
• Magnetic Resonance Spectroscopy 33%
• Human Brain Tumor 33%
• Proposed Methodology 33%
• Non-negative Matrix Factorization 33%
• Tumor Classification 33%
• Wide Applicability 33%
• Specific Information 33%
• Brain Tissue 33%
• Class-specific 33%
• Imaging Studies 33%
• Tissue Type 33%
• Non-invasive Imaging 33%
• Latent Variable Space 33%
• Source Extraction 33%
• Tumor Extent 33%
• Clinical Investigation 33%
• Factorization Method 33%
• Tumor Location 33%
• Multicenter Database 33%
• Single-voxel Spectroscopy 33%
• Mean Spectrum 33%
• Semi-supervised Learning 33%
• Biomedical Signal Processing 33%
• Class Prototype 33%
• Magnetic Resonance Spectrum 33%
• Perfect Correlation 33%
• Signal Information 33%
• Tissue Spectroscopy 33%
• Convex Nonnegative Matrix Factorization 33%
• Tumor Labeling 33%
• Signals Sources Identification 33%
• Metabolic Fingerprinting 33%
• Source Identification 33%

Computer Science

• Class Information 100%
• nonnegative matrix factorization 100%
• Matrix Factorization 100%
• Source Extraction 50%
• Biomedical Signal Processing 50%
• Frequency Domain 50%
• Information Signal 50%
• Semisupervised Learning 50%

Engineering

• Matrix Factorization 100%
• Human Brain 100%
• Single Voxel 75%
• Applicability 25%
• Frequency Domain 25%
• Metrics 25%
• Source Signal 25%
• Information Signal 25%
• Clinical Investigation 25%
• Brain Tissue 25%
• Biomedical Signal Processing 25%