vargar/Non_binary_

1import streamlit as st2import tensorflow as tf3import numpy as np4from PIL import Image5 6# Load the model7model = tf.keras.models.load_model("vgg19_binary_nonbinary.h5")8 9def preprocess_image(image):10    # Convert RGBA to RGB if the image has an alpha channel11    if image.mode == "RGBA":12        image = image.convert("RGB")13    # Resize and normalize the image14    image = image.resize((224, 224))  # Resize to match model input size15    image = np.array(image) / 255.0   # Normalize pixel values16    image = np.expand_dims(image, axis=0)  # Add batch dimension17    return image18    19# Streamlit app20st.title("Binary vs Non-Binary Image Classification")21st.write("Upload an image to classify it as 'binary' or 'non-binary'.")22 23# File uploader24uploaded_file = st.file_uploader("Choose an image...", type=["jpg", "jpeg", "png"])25if uploaded_file is not None:26    # Display the uploaded image27    image = Image.open(uploaded_file)28    st.image(image, caption="Uploaded Image", use_column_width=True)29    st.write("Classifying...")30 31    # Preprocess and predict32    processed_image = preprocess_image(image)33    predictions = model.predict(processed_image)34    class_names = ["binary", "non-binary"]35    confidence = {class_names[i]: float(predictions[0][i]) for i in range(2)}36 37    # Display the prediction38    st.write("Prediction:")39    st.write(confidence)