app.py

import streamlit as st
import tensorflow as tf
from PIL import Image
from src.utils import *
from src.vertex import *

st.set_page_config(
    page_title="Hip-Implant Image Classification",
    page_icon=":robot:",
    layout="centered",
    initial_sidebar_state="expanded",
    menu_items={
        'How to use': "# Upload an image of a hip-implant (search <loose hip implant> on google), the app will classify the hip-implant as loose or in-control."
    }
)

#creating session states
create_session_state()



image = Image.open('./image/title.jpg')
st.image(image)
st.title(":red[My AI Journey] :blue[Nishant Guvvada] X-ray Assistant")

with st.sidebar:
    image = Image.open('./image/sidebar_image.jpg')
    st.image(image)
    st.markdown("<h2 style='text-align: center; color: red;'>Settings Tab</h2>", unsafe_allow_html=True)


    st.write("Model Settings:")

    #define the temeperature for the model
    temperature_value = st.slider('Temperature :', 0.0, 1.0, 0.2)
    st.session_state['temperature'] = temperature_value

    #define the temeperature for the model
    token_limit_value = st.slider('Token limit :', 1, 1024, 256)
    st.session_state['token_limit'] = token_limit_value

    #define the temeperature for the model
    top_k_value = st.slider('Top-K  :', 1,40,40)
    st.session_state['top_k'] = top_k_value

    #define the temeperature for the model
    top_p_value = st.slider('Top-P :', 0.0, 1.0, 0.8)
    st.session_state['top_p'] = top_p_value

    if st.button("Reset Session"):
        reset_session()


st.image(bytes_data, caption='User uploaded image')
st.balloons()


@st.cache(allow_output_mutation=True)
def load_model():
    model=tf.keras.models.load_model('/Hip-Implant/hip_impant_model.h5')
    return model
with st.spinner('Model is being loaded..'):
    model=load_model()

st.write("""
         # Image Classification
         """
         )

file = st.file_uploader("Upload an X-ray image")
import cv2
from PIL import Image, ImageOps
import numpy as np
st.set_option('deprecation.showfileUploaderEncoding', False)
def model_prediction(img, model):
    resize = tf.image.resize(img, (256,256))
    yhat = model.predict(np.expand_dims(resize/255, 0))
    if(yhat>0.5):
        result = "Prediction is loose"
    else:
        result = "Prediction is control"
    return result
        
if file is None:
    st.text("Please upload an image file")
else:
    image = Image.open(file)
    st.image(image, use_column_width=True)
    predictions = mode_prediction(image, model)
    st.write(prediction)
    print(
    "This image most likely belongs to {}."
    .format(prediction)
)