Nabeelp/Image_Text_Classifier

1import streamlit as st2import numpy as np3from PIL import Image4import tensorflow5from tensorflow.keras.models import load_model6from tensorflow.keras.datasets import imdb7from tensorflow.keras.preprocessing.sequence import pad_sequences8import pickle9import sys10 11sys.path.append(r"E:\np\IMPORTANT\Stremlit(DL)\perceptron")12sys.path.append(r"E:\np\IMPORTANT\Stremlit(DL)\Back_propagation")13 14# Load word index for Sentiment Classification15word_to_index = imdb.get_word_index()16 17# Function to perform sentiment classification18def sentiment_classification(new_review_text, model):19    max_review_length = 50020    new_review_tokens = [word_to_index.get(word, 0) for word in new_review_text.split()]21    new_review_tokens = pad_sequences([new_review_tokens], maxlen=max_review_length)22    prediction = model.predict(new_review_tokens)23    if type(prediction) == list:24        prediction = prediction[0]25    return "Positive" if prediction > 0.5 else "Negative"26 27# Function to perform tumor detection28def tumor_detection(img, model):29    img = Image.open(img)30    img=img.resize((128,128))31    img=np.array(img)32    input_img = np.expand_dims(img, axis=0)33    res = model.predict(input_img)34    return "Tumor Detected" if res else "No Tumor"35 36# Streamlit App37st.title("Deep Prediction Hub")38 39# Choose between tasks40task = st.radio("Select Task", ("Sentiment Classification", "Tumor Detection"))41 42if task == "Sentiment Classification":43    # Input box for new review44    new_review_text = st.text_area("Enter a New Review:", value="")45    if st.button("Submit") and not new_review_text.strip():46        st.warning("Please enter a review.")47 48    if new_review_text.strip():49        st.subheader("Choose Model for Sentiment Classification")50        model_option = st.selectbox("Select Model", ("Perceptron", "Backpropagation", "DNN", "RNN", "LSTM"))51 52        # Load models dynamically based on the selected option53        if model_option == "Perceptron":54            with open('imdb_perceptron.pkl', 'rb') as file:55                model = pickle.load(file)56        elif model_option == "Backpropagation":57            with open('imdb_back_prop.pkl', 'rb') as file:58                model = pickle.load(file)59        elif model_option == "DNN":60            model = load_model('imdb_DNN.h5')61        elif model_option == "RNN":62            model = load_model('imdb_RNN.h5')63        elif model_option == "LSTM":64            model = load_model('lstm_model.h5')65 66        if st.button("Classify Sentiment"):67            result = sentiment_classification(new_review_text, model)68            st.subheader("Sentiment Classification Result")69            st.write(f"**{result}**")70 71elif task == "Tumor Detection":72    st.subheader("Tumor Detection")73    uploaded_file = st.file_uploader("Choose a tumor image...", type=["jpg", "jpeg", "png"])74 75    if uploaded_file is not None:76        # Load the tumor detection model77        model = load_model('CNN_.h5')78        st.image(uploaded_file, caption="Uploaded Image.", use_column_width=False, width=200)79        st.write("")80 81        if st.button("Detect Tumor"):82            result = tumor_detection(uploaded_file, model)83            st.subheader("Tumor Detection Result")84            st.write(f"**{result}**")