|
import streamlit as st |
|
import numpy as np |
|
from PIL import Image |
|
from tensorflow.keras.models import load_model |
|
from tensorflow.keras.preprocessing.text import Tokenizer |
|
from tensorflow.keras.preprocessing.sequence import pad_sequences |
|
from tensorflow.keras.applications.inception_v3 import preprocess_input |
|
import tensorflow as tf |
|
import joblib |
|
|
|
|
|
image_model = load_model('tumor_detection_model.h5') |
|
dnn_model = load_model('sms_spam_detection_dnnmodel.h5') |
|
rnn_model = load_model('spam_detection_rnn_model.h5') |
|
perceptron_model = joblib.load('imdb_perceptron_model.pkl') |
|
backprop_model = joblib.load('backprop_model.pkl') |
|
LSTM_model = load_model('imdb_LSTM.h5') |
|
|
|
|
|
st.title("Classification") |
|
|
|
|
|
task = st.sidebar.selectbox("Select Task", ["Tumor Detection", "Sentiment Classification"]) |
|
|
|
def preprocess_message_dnn(message, tokeniser, max_length): |
|
encoded_message = tokeniser.texts_to_sequences([message]) |
|
padded_message = pad_sequences(encoded_message, maxlen=max_length, padding='post') |
|
return padded_message |
|
|
|
def predict_dnnspam(message, tokeniser, max_length): |
|
processed_message = preprocess_message_dnn(message, tokeniser, max_length) |
|
prediction = dnn_model.predict(processed_message) |
|
return "Spam" if prediction >= 0.5 else "Ham" |
|
|
|
|
|
|
|
|
|
def preprocess_image(image): |
|
image = image.resize((299, 299)) |
|
image_array = np.array(image) |
|
preprocessed_image = preprocess_input(image_array) |
|
return preprocessed_image |
|
|
|
def make_prediction_cnn(image, model): |
|
img = image.resize((128, 128)) |
|
img_array = np.array(img) |
|
img_array = img_array.reshape((1, img_array.shape[0], img_array.shape[1], img_array.shape[2])) |
|
preprocessed_image = preprocess_input(img_array) |
|
prediction = model.predict(preprocessed_image) |
|
return "Tumor Detected" if prediction > 0.5 else "No Tumor" |
|
|
|
if task == "Sentiment Classification": |
|
st.subheader("Choose Model") |
|
model_choice = st.radio("Select Model", ["DNN", "RNN", "Perceptron", "Backpropagation", "LSTM"]) |
|
|
|
st.subheader("Text Input") |
|
text_input = st.text_area("Enter Text") |
|
|
|
if st.button("Predict"): |
|
if model_choice == "DNN": |
|
|
|
prediction_result = predict_dnnspam(text_input, tokeniser, max_length) |
|
st.write(f"The message is classified as: {prediction_result}") |
|
|
|
|
|
else: |
|
st.subheader("Choose Model") |
|
model_choice = st.radio("Select Model", ["CNN"]) |
|
|
|
st.subheader("Image Input") |
|
image_input = st.file_uploader("Choose an image...", type="jpg") |
|
|
|
if image_input is not None: |
|
image = Image.open(image_input) |
|
st.image(image, caption="Uploaded Image.", use_column_width=True) |
|
|
|
if st.button("Predict"): |
|
if model_choice == "CNN": |
|
prediction_result = make_prediction_cnn(image, image_model) |
|
st.write(prediction_result) |
|
|