Mandarin VS Japnense OR ANY LANGUAGE Image Classifier

# Issue with collecting meaningfull data which can be usefull
Using search api is providing data with certain constraints:
- Non Uniformity of dataset - Scaling of images
- Mixed Language text is being collected
- optimizing the size of the image to small leads to loss of details for images with large amount of text