This repo is based on three repositories below:
-
This repo is inspired by https://github.com/dotnet-labs/HandwritingRecognition I saw when I looked for many documents about OCR.
-
I created an image classifier model with MNIST dataset in JPG format. You can read more about the method at https://github.com/dotnet/machinelearning-samples/tree/main/samples/csharp/getting-started/MulticlassClassification_MNIST. However, I found that the results are not better than SDCA method with same image.
- This is a multiclass classification model using SDCA (Stochastic Dual Coordinate Ascent) algorithm https://github.com/dotnet/machinelearning-samples/tree/main/samples/csharp/getting-started/DeepLearning_ImageClassification_Training.