Cloud-PACS-AI-Integration / app-gradio.py
Kalbe-x-Bangkit's picture
Upload 13 files
fcc071b verified
raw
history blame
7.35 kB
import streamlit as st
import cv2
import numpy as np
import tensorflow as tf
from tensorflow.keras.preprocessing import image
from tensorflow.keras.models import load_model
import pandas as pd
class GradCAM:
def __init__(self, model, layer_name):
self.model = model
self.layer_name = layer_name
self.grad_model = tf.keras.models.Model(
[self.model.inputs],
[self.model.get_layer(layer_name).output, self.model.output]
)
def __call__(self, img_array, cls):
with tf.GradientTape() as tape:
conv_outputs, predictions = self.grad_model(img_array)
loss = predictions[:, cls]
output = conv_outputs[0]
grads = tape.gradient(loss, conv_outputs)[0]
gate_f = tf.cast(output > 0, 'float32')
gate_r = tf.cast(grads > 0, 'float32')
guided_grads = gate_f * gate_r * grads
weights = tf.reduce_mean(guided_grads, axis=(0, 1))
cam = np.zeros(output.shape[0:2], dtype=np.float32)
for index, w in enumerate(weights):
cam += w * output[:, :, index]
cam = cv2.resize(cam.numpy(), (224, 224))
cam = np.maximum(cam, 0)
cam = cam / cam.max()
return cam
def apply_heatmap(img, heatmap, heatmap_ratio=0.6):
heatmap = cv2.applyColorMap(np.uint8(255 * heatmap), cv2.COLORMAP_JET)
heatmap = cv2.cvtColor(heatmap, cv2.COLOR_BGR2RGB)
return np.uint8(heatmap * heatmap_ratio + img * (1 - heatmap_ratio))
def load_image(img_path, df, preprocess=True, H=320, W=320):
mean, std = get_mean_std_per_batch(img_path, df, H=H, W=W)
x = image.load_img(img_path, target_size=(H, W))
x = image.img_to_array(x)
if preprocess:
x -= mean
x /= std
x = np.expand_dims(x, axis=0)
return x
def get_mean_std_per_batch(image_path, df, H=320, W=320):
sample_data = []
for idx, img in enumerate(df.sample(100)["Image Index"].values):
sample_data.append(
np.array(image.load_img(image_path, target_size=(H, W))))
mean = np.mean(sample_data[0])
std = np.std(sample_data[0])
return mean, std
def compute_gradcam(img, model, df, labels, layer_name='bn'):
preprocessed_input = load_image(img, df)
predictions = model.predict(preprocessed_input)
top_indices = np.argsort(predictions[0])[-3:][::-1]
top_labels = [labels[i] for i in top_indices]
top_predictions = [predictions[0][i] for i in top_indices]
original_image = load_image(img, df, preprocess=False)
grad_cam = GradCAM(model, layer_name)
gradcam_images = []
for i in range(3):
idx = top_indices[i]
label = top_labels[i]
prob = top_predictions[i]
gradcam = grad_cam(preprocessed_input, idx)
gradcam_image = apply_heatmap(original_image, gradcam)
gradcam_images.append((gradcam_image, f"{label}: p={prob:.3f}"))
return gradcam_images
def calculate_mse(original_image, enhanced_image):
mse = np.mean((original_image - enhanced_image) ** 2)
return mse
def calculate_psnr(original_image, enhanced_image):
mse = calculate_mse(original_image, enhanced_image)
if mse == 0:
return float('inf')
max_pixel_value = 255.0
psnr = 20 * np.log10(max_pixel_value / np.sqrt(mse))
return psnr
def calculate_maxerr(original_image, enhanced_image):
maxerr = np.max((original_image - enhanced_image) ** 2)
return maxerr
def calculate_l2rat(original_image, enhanced_image):
l2norm_ratio = np.sum(original_image ** 2) / np.sum((original_image - enhanced_image) ** 2)
return l2norm_ratio
def process_image(original_image, enhancement_type, fix_monochrome=True):
if fix_monochrome and original_image.shape[-1] == 3:
original_image = cv2.cvtColor(original_image, cv2.COLOR_BGR2GRAY)
image = original_image - np.min(original_image)
image = image / np.max(original_image)
image = (image * 255).astype(np.uint8)
enhanced_image = enhance_image(image, enhancement_type)
mse = calculate_mse(original_image, enhanced_image)
psnr = calculate_psnr(original_image, enhanced_image)
maxerr = calculate_maxerr(original_image, enhanced_image)
l2rat = calculate_l2rat(original_image, enhanced_image)
return enhanced_image, mse, psnr, maxerr, l2rat
def apply_clahe(image):
clahe = cv2.createCLAHE(clipLimit=40.0, tileGridSize=(8, 8))
return clahe.apply(image)
def invert(image):
return cv2.bitwise_not(image)
def hp_filter(image, kernel=None):
if kernel is None:
kernel = np.array([[-1, -1, -1], [-1, 9, -1], [-1, -1, -1]])
return cv2.filter2D(image, -1, kernel)
def unsharp_mask(image, radius=5, amount=2):
def usm(image, radius, amount):
blurred = cv2.GaussianBlur(image, (0, 0), radius)
sharpened = cv2.addWeighted(image, 1.0 + amount, blurred, -amount, 0)
return sharpened
return usm(image, radius, amount)
def hist_eq(image):
return cv2.equalizeHist(image)
def enhance_image(image, enhancement_type):
if enhancement_type == "Invert":
return invert(image)
elif enhancement_type == "High Pass Filter":
return hp_filter(image)
elif enhancement_type == "Unsharp Masking":
return unsharp_mask(image)
elif enhancement_type == "Histogram Equalization":
return hist_eq(image)
elif enhancement_type == "CLAHE":
return apply_clahe(image)
else:
raise ValueError(f"Unknown enhancement type: {enhancement_type}")
st.title("Image Enhancement and Quality Evaluation")
uploaded_file = st.file_uploader("Upload Original Image", type=["png", "jpg", "jpeg"])
enhancement_type = st.radio("Enhancement Type", ["Invert", "High Pass Filter", "Unsharp Masking", "Histogram Equalization", "CLAHE"])
if uploaded_file is not None:
original_image = np.array(image.load_img(uploaded_file, color_mode='rgb' if enhancement_type == "Invert" else 'grayscale'))
enhanced_image, mse, psnr, maxerr, l2rat = process_image(original_image, enhancement_type)
st.image(original_image, caption='Original Image', use_column_width=True)
st.image(enhanced_image, caption='Enhanced Image', use_column_width=True)
st.write("MSE:", mse)
st.write("PSNR:", psnr)
st.write("Maxerr:", maxerr)
st.write("L2Rat:", l2rat)
st.title("Grad-CAM Visualization")
uploaded_gradcam_file = st.file_uploader("Upload Image for Grad-CAM", type=["png", "jpg", "jpeg"], key="gradcam")
if uploaded_gradcam_file is not None:
df_file = st.file_uploader("Upload DataFrame for Mean/Std Calculation", type=["csv"])
labels = st.text_area("Labels", placeholder="Enter labels separated by commas")
model_path = st.text_input("Model Path", 'model/densenet.hdf5')
pretrained_model_path = st.text_input("Pretrained Model Path", 'model/pretrained_model.h5')
if df_file and labels and model_path and pretrained_model_path:
df = pd.read_csv(df_file)
labels = labels.split(',')
model = load_model(model_path)
pretrained_model = load_model(pretrained_model_path)
gradcam_images = compute_gradcam(uploaded_gradcam_file, pretrained_model, df, labels)
for idx, (gradcam_image, label) in enumerate(gradcam_images):
st.image(gradcam_image, caption=f'Grad-CAM {idx+1}: {label}', use_column_width=True)