Implementation of a simple image segmentation algorithm of a natural scene image by taking its Laplacian of Gaussian in frequency domain and converting it back to time domain to get a spatially-enhanced image.
Next applying K-means clustering algorithm with n=2 to separate out text and background of the image.
collage.png show the transformtation of the iamge.
Run the text_clustering.py file on any natural scene image with text in it to generate the text and non-text clusters for it.