image_filtering.py

"""
The code here hepls at cleaning grayscale image to
whitten image and darken other parts
"""
import os
import numpy as np
import skimage
import pathlib
from os import listdir
from os.path import isfile
from tkinter import *
from tkinter import Tcl
import skimage.io as io
from utils.config import PROJECT_ROOT
from scipy import ndimage
from utils.config import roi_image
from utils.make_dir import create_dir

# Get the project root directory
project_path = PROJECT_ROOT
os.chdir(PROJECT_ROOT)


def whitten_image_darken_else_filtering(input_img_path, save_path):
    """
    This function is used to filter the image directly after it is out from the deep learning
    """
    nlyTIFF = [
        os.path.join(input_img_path, f)
        for f in listdir(input_img_path)
        if isfile(os.path.join(input_img_path, f)) and f.endswith(".jpg")
    ]
    nlyTIFF = Tcl().call("lsort", "-dict", nlyTIFF)
    print(input_img_path)
    if len(nlyTIFF) >= 2:
        for i in range(len(nlyTIFF)):
            grayscale = io.imread(nlyTIFF[i], plugin="matplotlib")
            median_filtered = ndimage.median_filter(grayscale, size=3)
            threshold = skimage.filters.threshold_li(median_filtered)
            predicted = np.uint8(median_filtered > threshold) * 255
            io.imsave(
                os.path.join(
                    save_path, "{}.{}".format(input_img_path.split("/")[-1], "jpg")
                ),
                predicted,
            )

    elif len(nlyTIFF) == 1:
        from PIL import Image

        Image.MAX_IMAGE_PIXELS = 1000000000
        grayscale = io.imread(nlyTIFF[0], plugin="matplotlib")
        median_filtered = ndimage.median_filter(grayscale, size=3)
        threshold = skimage.filters.threshold_li(grayscale)
        predicted = np.uint8(median_filtered > threshold) * 255
        io.imsave(
            os.path.join(
                save_path, "{}.{}".format(input_img_path.split("/")[-1], "jpg")
            ),
            predicted,
        )
    else:
        print("No jpg file given in the path provided")


new_tif = roi_image.split(".")[0]
input_path = PROJECT_ROOT + "results/Test/predicted/" + new_tif

# Path to save the outputs
save_path = PROJECT_ROOT + "results/Test/filtered/"
# Create dir for saving predictions
output_dir = create_dir(save_path + roi_image.split(".")[0])
save_path = pathlib.Path(output_dir)

# Function call
whitten_image_darken_else_filtering(input_path, save_path)