Deep residual convolutional neural network with curriculum learning for source camera identification

Keyphrases

• Curriculum Learning 100%
• Source Camera Identification 100%
• Deep Residual Convolutional Neural Network 100%
• U-Net 75%
• Camera Identification 50%
• Convolutional Layer 50%
• State-of-the-art Techniques 25%
• Learning Algorithm 25%
• Individual-based 25%
• Discriminative Features 25%
• Identification Accuracy 25%
• One-vs-One 25%
• Linear Support Vector Machine 25%
• Deep Learning Methods 25%
• Existing State 25%
• Query Image 25%
• Forensic Science 25%
• Camera Image 25%
• Camera Model 25%
• Noise Residue 25%
• Dresden 25%
• Forensic Analysis 25%
• Instance-based 25%
• Fully Connected Layer 25%
• Residual Connection 25%
• Automatic Curriculum Learning 25%
• Camera Fingerprint 25%
• Child Pornography 25%
• Camera Processing pipeline 25%
• Small-sized 25%

Computer Science

• Convolutional Neural Network 100%
• Residual Neural Network 100%
• Convolutional Layer 66%
• Deep Learning Method 33%
• Discriminative Feature 33%
• Deep Feature 33%
• Support Vector Machine 33%
• Learning Algorithm 33%
• Experimental Result 33%
• Fully Connected Layer 33%