Cutting the Error by Half: Investigation of Very Deep CNN and Advanced Training Strategies for Document Image Classification.
Muhammad Zeshan AfzalAndreas KölschSheraz AhmedMarcus LiwickiPublished in: ICDAR (2017)
Keyphrases
- image classification
- classifier training
- bag of words
- cellular neural networks
- image representation
- document collections
- information retrieval
- training set
- visual features
- error bounds
- error rate
- document retrieval
- text documents
- retrieval strategies
- document classification
- visual words
- feature extraction
- multi label
- retrieval systems
- web documents
- document clustering
- training examples
- supervised learning
- training data
- sparse representation
- training samples
- training process
- structured documents
- document analysis
- image features
- centroid neural network
- total error rate