Performance Analysis of SA-AMG Method by Setting Extracted Near-Kernel Vectors.
Naoya NomuraAkihiro FujiiTeruo TanakaKengo NakajimaOsni MarquesPublished in: VECPAR (2016)
Keyphrases
- detection method
- high accuracy
- computational cost
- experimental evaluation
- principal components
- high precision
- significant improvement
- cost function
- dynamic programming
- prior knowledge
- support vector machine svm
- synthetic data
- objective function
- support vector
- decision trees
- artificial neural networks
- edge detection
- preprocessing
- kernel function