i am working on a camera based document image analysis, where i have to identify the script from multi-lingual document images, hence i need to extract the features using alexnet is it suitable to extract the document image features.
Thank you Luay Shihab, for valuable information regarding the CNN, yes the CNN is one of the important field in deep learning and it has got lots of pretrained CNNs i knew the concept but i need to implement the CNN on my dataset. My objective is to identify the script from different document images for that i have created the dataset of 24,000 images of having 6 classes, from this dataset i need to extract the features and want train them using SVM multiclass. i need some detailed implementation information for CNN.