File size: 5,788 Bytes
6860cad |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 |
import os
# Silence TF logs
os.environ.setdefault("TF_CPP_MIN_LOG_LEVEL", "2")
os.environ.setdefault("TF_ENABLE_ONEDNN_OPTS", "0")
import gradio as gr
import numpy as np
import cv2
from tensorflow.keras.models import load_model
from skimage.metrics import structural_similarity as ssim
# =======================
# Load trained model
# =======================
MODEL_PATH = "image_denoising_autoencoder.h5"
model = load_model(MODEL_PATH, compile=False)
# =======================
# Utility Functions
# =======================
def preprocess_gray(image):
image = cv2.resize(image, (128, 128))
image = image.astype("float32") / 255.0
return image.reshape(128, 128, 1)
def add_noise(image, noise_factor):
noisy = image + noise_factor * np.random.normal(
loc=0.0, scale=1.0, size=image.shape
)
return np.clip(noisy, 0.0, 1.0)
def psnr(original, denoised):
mse = np.mean((original - denoised) ** 2)
if mse == 0:
return 100
return 10 * np.log10(1.0 / mse)
def interpret_psnr(value):
if value < 15:
return "Poor ❌"
elif value < 25:
return "Moderate ⚠️"
else:
return "Good ✅"
def interpret_ssim(value):
if value < 0.5:
return "Low ❌"
elif value < 0.8:
return "Moderate ⚠️"
else:
return "High ✅"
# =======================
# 🔥 POST-PROCESSING
# =======================
def enhance_clarity(image, strength=1.0):
image_uint8 = (image * 255).astype(np.uint8)
blurred = cv2.GaussianBlur(image_uint8, (5, 5), 0)
sharpened = cv2.addWeighted(
image_uint8, 1 + strength, blurred, -strength, 0
)
return np.clip(sharpened.astype("float32") / 255.0, 0, 1)
# =======================
# Inference Function
# =======================
def denoise_image(input_image, noise_level, mode, clarity_strength):
if input_image is None:
return None, None, None, "Upload an image to start."
# -------- Grayscale --------
if mode == "Grayscale":
if input_image.ndim == 3:
gray = cv2.cvtColor(input_image, cv2.COLOR_RGB2GRAY)
else:
gray = input_image
clean = preprocess_gray(gray)
noisy = add_noise(clean, noise_level)
denoised = model.predict(noisy.reshape(1, 128, 128, 1))[0]
enhanced = enhance_clarity(
denoised.reshape(128, 128),
strength=clarity_strength
)
psnr_val = psnr(clean.reshape(128, 128), denoised.reshape(128, 128))
ssim_val = ssim(
clean.reshape(128, 128),
denoised.reshape(128, 128),
data_range=1.0
)
metrics = (
f"PSNR: {psnr_val:.2f} dB ({interpret_psnr(psnr_val)})\n"
f"SSIM: {ssim_val:.3f} ({interpret_ssim(ssim_val)})\n"
f"Clarity Strength: {clarity_strength:.2f}"
)
return (
clean.reshape(128, 128),
noisy.reshape(128, 128),
enhanced,
metrics
)
# -------- RGB --------
resized = cv2.resize(input_image, (128, 128))
resized = resized.astype("float32") / 255.0
noisy_rgb = add_noise(resized, noise_level)
denoised_rgb = np.zeros_like(noisy_rgb)
for c in range(3):
channel = noisy_rgb[:, :, c].reshape(128, 128, 1)
denoised_rgb[:, :, c] = model.predict(
channel.reshape(1, 128, 128, 1)
)[0].reshape(128, 128)
enhanced_rgb = np.zeros_like(denoised_rgb)
for c in range(3):
enhanced_rgb[:, :, c] = enhance_clarity(
denoised_rgb[:, :, c],
strength=clarity_strength
)
psnr_val = psnr(resized, denoised_rgb)
ssim_val = ssim(
resized,
denoised_rgb,
channel_axis=2,
data_range=1.0
)
metrics = (
f"PSNR: {psnr_val:.2f} dB ({interpret_psnr(psnr_val)})\n"
f"SSIM: {ssim_val:.3f} ({interpret_ssim(ssim_val)})\n"
f"Clarity Strength: {clarity_strength:.2f}"
)
return resized, noisy_rgb, enhanced_rgb, metrics
# =======================
# Gradio UI
# =======================
with gr.Blocks(title="Image Denoising Autoencoder") as demo:
gr.Markdown(
"""
# 🧠 Image Denoising using U-Net Autoencoder
🎚 Sliders update output automatically
"""
)
with gr.Row():
input_image = gr.Image(label="Upload Image", type="numpy", height=320)
with gr.Column():
noise_slider = gr.Slider(0.05, 0.5, 0.3, 0.05, label="Noise Level")
clarity_slider = gr.Slider(0.0, 2.0, 1.0, 0.1, label="Clarity Strength")
mode_selector = gr.Radio(
["Grayscale", "RGB"], value="Grayscale", label="Mode"
)
denoise_btn = gr.Button("Denoise Image 🚀")
with gr.Row():
original_out = gr.Image(label="Original", height=260)
noisy_out = gr.Image(label="Noisy", height=260)
denoised_out = gr.Image(label="Denoised", height=260)
metrics_box = gr.Textbox(label="Quality Metrics", lines=4)
inputs = [input_image, noise_slider, mode_selector, clarity_slider]
outputs = [original_out, noisy_out, denoised_out, metrics_box]
# 🔁 AUTO UPDATE
noise_slider.change(denoise_image, inputs, outputs)
clarity_slider.change(denoise_image, inputs, outputs)
mode_selector.change(denoise_image, inputs, outputs)
# 🔘 MANUAL BUTTON
denoise_btn.click(denoise_image, inputs, outputs)
# =======================
# Launch App
# =======================
if __name__ == "__main__":
demo.launch(debug=True)
|