app.py

import streamlit as st
from tensorflow.keras.models import load_model
from PIL import Image
import numpy as np
import matplotlib.pyplot as plt
import seaborn as sns

model = load_model('cnn_model.h5')

def process_image(img):
    img = img.convert('RGB')  
    img = img.resize((64, 64)) 
    img = np.array(img)  
    img = img / 255.0  
    img = np.expand_dims(img, axis=0) 
    return img

st.title('Fruits and Vegetables Image Recognition :grapes: 	:carrot: :kiwifruit: :broccoli:')
st.write('Upload a vegetable or fruit image and the model will recognize it.')

file = st.file_uploader('Select an image', type=['jpg', 'jpeg', 'png'])

if file is not None:
    # Display the uploaded image
    img = Image.open(file)
    st.image(img, caption='Uploaded Image')

    # Preprocess the image
    image = process_image(img)
    
    # Model prediction
    with st.spinner('Classifying the image...'):
        predictions = model.predict(image)
        predicted_class = np.argmax(predictions)  
        predicted_prob = predictions[0][predicted_class]  

    # Class names for prediction
    class_names = ['capsicum',
                'sweetcorn',
                'orange',
                'tomato',
                'turnip',
                'ginger',
                'raddish',
                'pomegranate',
                'pineapple',
                'jalepeno',
                'apple',
                'carrot',
                'lettuce',
                'bell pepper',
                'eggplant',
                'beetroot',
                'kiwi',
                'pear',
                'cabbage',
                'cauliflower',
                'paprika',
                'lemon',
                'sweetpotato',
                'grapes',
                'cucumber',
                'corn',
                'banana',
                'garlic',
                'chilli pepper',
                'watermelon',
                'mango',
                'peas',
                'onion',
                'potato',
                'spinach',
                'soy beans']
    
    # Display the prediction
    st.subheader(f"Prediction: {class_names[predicted_class]}")
    st.write(f"Confidence: {predicted_prob * 100:.2f}%")

    # Display prediction probabilities
    st.write("Prediction Probabilities for Each Class:")
    probabilities = predictions[0]
    prob_dict = {class_names[i]: probabilities[i] for i in range(len(class_names))}
    
    # Plot settings
    sns.set(style="white")  # Clean style with no grid background

    # Create the figure
    fig, ax = plt.subplots(figsize=(12, 8))  # Adjust figure size for better readability

    # Plot the bar chart
    ax.bar(list(prob_dict.keys()), list(prob_dict.values()), color='#197339', edgecolor='black')
    ax.set_ylabel('Probability', fontsize=14)
    ax.set_title('Prediction Probabilities for Each Class', fontsize=20)

    # Rotate x-axis labels for better readability
    plt.xticks(rotation=90, ha='right', fontsize=15)

    # Annotate bars with percentage values
    for index, value in enumerate(prob_dict.values()):
        ax.text(index, value, f'{value * 100:.0f}%', va='bottom', ha='center', fontsize=10)

    # Remove background grid and spines
    ax.spines['top'].set_visible(False)
    ax.spines['right'].set_visible(False)
    ax.spines['left'].set_visible(False)
    ax.spines['bottom'].set_visible(False)
    ax.grid(False)

    # Adjust layout to prevent clipping
    fig.tight_layout()

    # Display the plot in Streamlit
    st.pyplot(fig)