Aim: Detecting and Classifying Hieroglyphs using CNN and Logistic Regression Steps
-
Phase 1
- Reading dataset and performing label encoding on the labels
- Splitting the dataset into training and testing
- Using Inception V3 for extraing features
- Using these features to train and test using Logistic Regression
- Evaluating the model using accuracy
- Testing model on unseen data
-
Phase 2
- Read the dataset
- Augment the dataset
- Label encoding the labels
- Perform one hot encoding on the labels
- Split into training and testing dataset
- Develop CNN model
- Evaluate on
- Loss
- Accuracy
- Precision
- Recall
- F1 Score
- Model is generalised on another dataset
Conclusion
- CNN is better performing
- Model is generalised