Img.py

import streamlit as st
import numpy as np
from PIL import Image
import cv2
from tensorflow import keras

"""
   Using Streamlit to Predict the type of WBC that is present in the Image.   
"""

## Title for the Page
st.title("White Blood Cell Classification")

## Widget to Upload an Image
uploaded_file = st.file_uploader("Upload the Microscopic Image of the Blood Sample:",
                                 type=["jpeg","jpg", "png"])


if uploaded_file is not None:
    image_orig = Image.open(uploaded_file).convert('RGB')  # open and convert to RGB format
    st.image(image_orig, caption='Uploaded image', use_column_width=True)  # Displays the Image on to the page
    st.write("Classifying...")
    MODEL_PATH = 'E:\Microscopic Insights\Microscopic\Microscopic Images\Microscopic-Images-Insights\weights\model_2.h5'  # Path to the Model Weights
    model = keras.models.load_model(MODEL_PATH)  # Model Loaded Using Keras API
    my_bar = st.progress(0) # Progress bar set at 0

    img_array = np.array(image_orig) 
    img_array = cv2.resize(img_array, (320, 240), interpolation=cv2.INTER_NEAREST)  # norm
    my_bar.progress(25)  # Progress bar set at 25
    test_im = np.expand_dims(img_array, axis=0)

    my_bar.progress(50)  # Progress bar set at 50
    prediction = model.predict(test_im / 255)  # Model predicts the type of WBC and returns an Integer Value
    my_bar.progress(75)  # Progress bar set at 75
    prediction = np.argmax(prediction)
    if prediction == 0:
        string = "EOSINOPHIL"
    elif prediction == 1:
        string = "LYMPHOCYTE"
    elif prediction == 2:
        string = "MONOCYTE"
    elif prediction == 3:
        string = "NEUTROPHIL"
    st.write(f"WBC type :- {string}")

    my_bar.progress(100) # Progess bar reaches 100
    st.success('Done')