app.py

import gradio as gr
import numpy as np
import torch
import cv2
from PIL import Image
from torchvision import transforms
from cloth_segmentation.networks.u2net import U2NET
import matplotlib.colors as mcolors

# Load U²-Net
model_path = "cloth_segmentation/networks/u2net.pth"
model = U2NET(3, 1)
state_dict = torch.load(model_path, map_location=torch.device("cpu"))
state_dict = {k.replace("module.", ""): v for k, v in state_dict.items()}
model.load_state_dict(state_dict)
model.eval()

# Util to get BGR color from name
def get_bgr_from_color_name(color_name):
    try:
        rgb = mcolors.to_rgb(color_name.lower())
        return tuple(int(255 * c) for c in rgb[::-1])  # Convert to BGR
    except:
        return (0, 0, 255)  # Default to red

# Mask refinement
def refine_mask(mask):
    close_kernel = np.ones((5, 5), np.uint8)
    mask = cv2.morphologyEx(mask, cv2.MORPH_CLOSE, close_kernel)
    erode_kernel = np.ones((3, 3), np.uint8)
    mask = cv2.erode(mask, erode_kernel, iterations=1)
    mask = cv2.morphologyEx(mask, cv2.MORPH_CLOSE, close_kernel)
    return cv2.GaussianBlur(mask, (5, 5), 1.5)

# U²-Net segmentation
def segment_dress(image_np):
    transform_pipeline = transforms.Compose([
        transforms.ToTensor(),
        transforms.Resize((320, 320))
    ])
    image = Image.fromarray(image_np).convert("RGB")
    input_tensor = transform_pipeline(image).unsqueeze(0)

    with torch.no_grad():
        output = model(input_tensor)[0][0].squeeze().cpu().numpy()

    output = (output - output.min()) / (output.max() - output.min() + 1e-8)
    adaptive_thresh = np.mean(output) + 0.2
    dress_mask = (output > adaptive_thresh).astype(np.uint8) * 255
    return refine_mask(cv2.resize(dress_mask, (image_np.shape[1], image_np.shape[0]), interpolation=cv2.INTER_NEAREST))

# Optional GrabCut refinement
def apply_grabcut(image_np, dress_mask):
    bgd_model = np.zeros((1, 65), np.float64)
    fgd_model = np.zeros((1, 65), np.float64)
    mask = np.where(dress_mask > 0, cv2.GC_PR_FGD, cv2.GC_BGD).astype('uint8')
    coords = cv2.findNonZero(dress_mask)
    if coords is not None:
        x, y, w, h = cv2.boundingRect(coords)
        rect = (x, y, w, h)
        cv2.grabCut(image_np, mask, rect, bgd_model, fgd_model, 3, cv2.GC_INIT_WITH_MASK)
    refined = np.where((mask == cv2.GC_FGD) | (mask == cv2.GC_PR_FGD), 255, 0).astype("uint8")
    return refine_mask(refined)

# LAB color recoloring
def recolor_dress(image_np, dress_mask, target_color):
    target_color_lab = cv2.cvtColor(np.uint8([[target_color]]), cv2.COLOR_BGR2LAB)[0][0]
    img_lab = cv2.cvtColor(image_np, cv2.COLOR_RGB2LAB)

    dress_pixels = img_lab[dress_mask > 0]
    if len(dress_pixels) == 0:
        return image_np

    mean_L, mean_A, mean_B = np.mean(dress_pixels, axis=0)
    a_shift = target_color_lab[1] - mean_A
    b_shift = target_color_lab[2] - mean_B

    img_lab[..., 1] = np.clip(img_lab[..., 1] + (dress_mask / 255.0) * a_shift, 0, 255)
    img_lab[..., 2] = np.clip(img_lab[..., 2] + (dress_mask / 255.0) * b_shift, 0, 255)

    img_recolored = cv2.cvtColor(img_lab.astype(np.uint8), cv2.COLOR_LAB2RGB)
    feathered_mask = cv2.GaussianBlur(dress_mask, (21, 21), 7)
    lightness_mask = (img_lab[..., 0] / 255.0) ** 0.7
    adaptive_feather = (feathered_mask * lightness_mask).astype(np.uint8)

    return (image_np * (1 - adaptive_feather[..., None] / 255) + img_recolored * (adaptive_feather[..., None] / 255)).astype(np.uint8)

# Main function
def change_dress_color(img, color_prompt):
    if img is None or not color_prompt:
        return img

    img_np = np.array(img)
    target_bgr = get_bgr_from_color_name(color_prompt)

    try:
        dress_mask = segment_dress(img_np)
        if np.sum(dress_mask) < 1000:
            return img
        dress_mask = apply_grabcut(img_np, dress_mask)
        img_recolored = recolor_dress(img_np, dress_mask, target_bgr)
        return Image.fromarray(img_recolored)
    except Exception as e:
        print(f"Error: {e}")
        return img

# Gradio UI
with gr.Blocks() as demo:
    gr.Markdown("## 🎨 AI Dress Recolorer - Prompt Based")
    gr.Markdown("Upload an image and type a color (e.g., 'lavender', 'light green', 'royal blue').")

    with gr.Row():
        with gr.Column():
            input_image = gr.Image(type="pil", label="Upload Image")
            color_input = gr.Textbox(label="Enter Dress Color", placeholder="e.g. crimson, lavender, sky blue")
            recolor_btn = gr.Button("Apply New Color")

        with gr.Column():
            output_image = gr.Image(type="pil", label="Recolored Result")

    recolor_btn.click(fn=change_dress_color, inputs=[input_image, color_input], outputs=output_image)

if __name__ == "__main__":
    demo.launch()