from typing import Dict, Optional, Tuple, List from dataclasses import dataclass import os import sys proj_dir = os.path.dirname(os.path.dirname(os.path.abspath(__file__))) sys.path.append(os.path.join(proj_dir)) import time import cv2 import gradio as gr import numpy as np import torch import PIL from PIL import Image import rembg from rembg import remove rembg_session = rembg.new_session() from segment_anything import sam_model_registry, SamPredictor import craftsman from craftsman.systems.base import BaseSystem from craftsman.utils.config import ExperimentConfig, load_config parent_dir = os.path.dirname(os.path.dirname(os.path.abspath(__file__))) def check_input_image(input_image): if input_image is None: raise gr.Error("No image uploaded!") def load_model( ckpt_path: str, config_path: str, device = "cuda" ): cfg: ExperimentConfig cfg = load_config(config_path) if 'pretrained_model_name_or_path' not in cfg.system.condition_model or cfg.system.condition_model.pretrained_model_name_or_path is None: cfg.system.condition_model.config_path = config_path.replace("config.yaml", "clip_config.json") system: BaseSystem = craftsman.find(cfg.system_type)( cfg.system, ) print(f"Restoring states from the checkpoint path at {ckpt_path} with config {cfg}") system.load_state_dict(torch.load(ckpt_path, map_location=torch.device('cpu'))['state_dict']) system = system.to(device).eval() return system class RMBG(object): def __init__(self, device): sam = sam_model_registry["vit_h"](checkpoint=f"{parent_dir}/ckpts/SAM/sam_vit_h_4b8939.pth").to(device) self.predictor = SamPredictor(sam) def rmbg_sam(self, input_image): def _sam_segment(predictor, input_image, *bbox_coords): bbox = np.array(bbox_coords) image = np.asarray(input_image) start_time = time.time() predictor.set_image(image) masks_bbox, scores_bbox, logits_bbox = predictor.predict(box=bbox, multimask_output=True) print(f"SAM Time: {time.time() - start_time:.3f}s") out_image = np.zeros((image.shape[0], image.shape[1], 4), dtype=np.uint8) out_image[:, :, :3] = image out_image_bbox = out_image.copy() out_image_bbox[:, :, 3] = masks_bbox[-1].astype(np.uint8) * 255 torch.cuda.empty_cache() return Image.fromarray(out_image_bbox, mode='RGBA') RES = 1024 input_image.thumbnail([RES, RES], Image.Resampling.LANCZOS) image_rem = input_image.convert('RGBA') image_nobg = remove(image_rem, alpha_matting=True) arr = np.asarray(image_nobg)[:, :, -1] x_nonzero = np.nonzero(arr.sum(axis=0)) y_nonzero = np.nonzero(arr.sum(axis=1)) x_min = int(x_nonzero[0].min()) y_min = int(y_nonzero[0].min()) x_max = int(x_nonzero[0].max()) y_max = int(y_nonzero[0].max()) return _sam_segment(self.predictor, input_image.convert('RGB'), x_min, y_min, x_max, y_max) def rmbg_rembg(self, input_image): def _rembg_remove( image: PIL.Image.Image, rembg_session = None, force: bool = False, **rembg_kwargs, ) -> PIL.Image.Image: do_remove = True if image.mode == "RGBA" and image.getextrema()[3][0] < 255: # explain why current do not rm bg print("alhpa channl not enpty, skip remove background, using alpha channel as mask") background = Image.new("RGBA", image.size, (0, 0, 0, 0)) image = Image.alpha_composite(background, image) do_remove = False do_remove = do_remove or force if do_remove: image = rembg.remove(image, session=rembg_session, **rembg_kwargs) return image return _rembg_remove(input_image, rembg_session, force_remove=True) def run(self, rm_type, image, foreground_ratio, background_choice, backgroud_color): # image = cv2.resize(np.array(image), (crop_size, crop_size)) # image = Image.fromarray(image) if background_choice == "Alpha as mask": background = Image.new("RGBA", image.size, (backgroud_color[0], backgroud_color[1], backgroud_color[2], 0)) return Image.alpha_composite(background, image) elif "Remove" in background_choice: if rm_type.upper() == "SAM": image = self.rmbg_sam(image) elif rm_type.upper() == "REMBG": image = self.rmbg_rembg(image) else: return -1 image = do_resize_content(image, foreground_ratio) image = expand_to_square(image) # image = add_background(image, backgroud_color) # return image.convert("RGB") return image elif "Original" in background_choice: return image else: return -1 def do_resize_content(original_image: Image, scale_rate): # resize image content wile retain the original image size if scale_rate != 1: # Calculate the new size after rescaling new_size = tuple(int(dim * scale_rate) for dim in original_image.size) # Resize the image while maintaining the aspect ratio resized_image = original_image.resize(new_size) # Create a new image with the original size and black background padded_image = Image.new("RGBA", original_image.size, (0, 0, 0, 0)) paste_position = ((original_image.width - resized_image.width) // 2, (original_image.height - resized_image.height) // 2) padded_image.paste(resized_image, paste_position) return padded_image else: return original_image def expand2square(pil_img, background_color): width, height = pil_img.size if width == height: return pil_img elif width > height: result = Image.new(pil_img.mode, (width, width), background_color) result.paste(pil_img, (0, (width - height) // 2)) return result else: result = Image.new(pil_img.mode, (height, height), background_color) result.paste(pil_img, ((height - width) // 2, 0)) return result def expand_to_square(image, bg_color=(0, 0, 0, 0)): # expand image to 1:1 width, height = image.size if width == height: return image new_size = (max(width, height), max(width, height)) new_image = Image.new("RGBA", new_size, bg_color) paste_position = ((new_size[0] - width) // 2, (new_size[1] - height) // 2) new_image.paste(image, paste_position) return new_image def add_background(image, bg_color=(255, 255, 255)): # given an RGBA image, alpha channel is used as mask to add background color background = Image.new("RGBA", image.size, bg_color) return Image.alpha_composite(background, image)