app.py

#!/usr/bin/env python
# coding: utf-8

# In[ ]:


import streamlit as st
import joblib
import json
import numpy as np
import base64
import cv2
import pywt
from PIL import Image
import io

__class_name_to_number = {}
__class_number_to_name = {}

__model = None

def w2d(img, mode='haar', level=1):
    imArray = img
    #Datatype conversions
    #convert to grayscale
    imArray = cv2.cvtColor( imArray,cv2.COLOR_RGB2GRAY )
    #convert to float
    imArray =  np.float32(imArray)
    imArray /= 255;
    # compute coefficients
    coeffs=pywt.wavedec2(imArray, mode, level=level)

    #Process Coefficients
    coeffs_H=list(coeffs)
    coeffs_H[0] *= 0;

    # reconstruction
    imArray_H=pywt.waverec2(coeffs_H, mode);
    imArray_H *= 255;
    imArray_H =  np.uint8(imArray_H)

    return imArray_H

def classify_image(image_base64_data, file_path=None):

    imgs = get_cropped_image_if_2_eyes(file_path, image_base64_data)

    result = []
    for img in imgs:
        scalled_raw_img = cv2.resize(img, (32, 32))
        img_har = w2d(img, 'db1', 5)
        scalled_img_har = cv2.resize(img_har, (32, 32))
        combined_img = np.vstack((scalled_raw_img.reshape(32 * 32 * 3, 1), scalled_img_har.reshape(32 * 32, 1)))

        len_image_array = 32*32*3 + 32*32

        final = combined_img.reshape(1,len_image_array).astype(float)
        result.append({
            'class': class_number_to_name(__model.predict(final)[0]),
            'class_probability': np.around(__model.predict_proba(final)*100,2).tolist()[0],
            'class_dictionary': __class_name_to_number
        })

    return result

def class_number_to_name(class_num):
    return __class_number_to_name[class_num]

def load_saved_artifacts():
    print("loading saved artifacts...start")
    global __class_name_to_number
    global __class_number_to_name

    with open("./class_dictionary.json", "r") as f:
        __class_name_to_number = json.load(f)
        __class_number_to_name = {v:k for k,v in __class_name_to_number.items()}

    global __model
    if __model is None:
        with open('./saved_model.pkl', 'rb') as f:
            __model = joblib.load(f)
    print("loading saved artifacts...done")


def get_cv2_image_from_base64_string(b64str):
    '''
    credit: https://stackoverflow.com/questions/33754935/read-a-base-64-encoded-image-from-memory-using-opencv-python-library
    :param uri:
    :return:
    '''
    encoded_data = b64str.split(',')[1]
    nparr = np.frombuffer(base64.b64decode(encoded_data), np.uint8)
    img = cv2.imdecode(nparr, cv2.IMREAD_COLOR)
    return img

def get_cropped_image_if_2_eyes(image_path, image_base64_data):
    face_cascade = cv2.CascadeClassifier('./opencv/haarcascades/haarcascade_frontalface_default.xml')
    eye_cascade = cv2.CascadeClassifier('./opencv/haarcascades/haarcascade_eye.xml')

    if image_path:
        img = cv2.imread(image_path)
    else:
        img = get_cv2_image_from_base64_string(image_base64_data)

    gray = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)
    faces = face_cascade.detectMultiScale(gray, 1.3, 5)

    cropped_faces = []
    for (x,y,w,h) in faces:
            roi_gray = gray[y:y+h, x:x+w]
            roi_color = img[y:y+h, x:x+w]
            eyes = eye_cascade.detectMultiScale(roi_gray)
            if len(eyes) >= 2:
                cropped_faces.append(roi_color)
    return cropped_faces

def get_b64_test_image_for_virat():
    with open("b64.txt") as f:
        return f.read()

# Load the saved artifacts at the start of the app
load_saved_artifacts()

st.title("Celebrity Image Classification")

uploaded_file = st.file_uploader("Choose an image...", type=["jpg", "jpeg", "png"])

if uploaded_file is not None:
    image = Image.open(uploaded_file)
    st.image(image, caption='Uploaded Image', use_column_width=True)
    
    # Convert PIL Image to base64
    buffered = io.BytesIO()
    image.save(buffered, format="PNG")
    img_str = base64.b64encode(buffered.getvalue()).decode()
    
    # Auto-classify the image
    result = classify_image(f"data:image/png;base64,{img_str}", None)
    
    if result:
        st.subheader("Classification Results:")
        for r in result:
            # Create a dictionary of celebrity names and their probabilities
            probabilities = dict(zip(r['class_dictionary'].keys(), r['class_probability']))
            
            # Sort the probabilities in descending order
            sorted_probabilities = sorted(probabilities.items(), key=lambda x: x[1], reverse=True)
            
            # Create a DataFrame for the tabulator
            import pandas as pd
            df = pd.DataFrame(sorted_probabilities, columns=['Celebrity', 'Probability'])
            
            # Display the tabulator
            st.table(df)
            
            # Display the top prediction
            st.write(f"Top prediction: {r['class']} with {r['class_probability'][r['class_dictionary'][r['class']]]}% probability")
    else:
        st.write("No faces detected in the image.")