from typing import Any, List, Literal, Optional from argparse import ArgumentParser from time import sleep import cv2 import numpy import onnxruntime import deepfuze.globals import deepfuze.processors.frame.core as frame_processors from deepfuze import config, process_manager, logger, wording from deepfuze.execution import apply_execution_provider_options from deepfuze.face_analyser import get_one_face, get_many_faces, find_similar_faces, clear_face_analyser from deepfuze.face_masker import create_static_box_mask, create_occlusion_mask, create_mouth_mask, clear_face_occluder, clear_face_parser from deepfuze.face_helper import warp_face_by_face_landmark_5, warp_face_by_bounding_box, paste_back, create_bounding_box_from_face_landmark_68 from deepfuze.face_store import get_reference_faces from deepfuze.content_analyser import clear_content_analyser from deepfuze.normalizer import normalize_output_path from deepfuze.thread_helper import thread_lock, conditional_thread_semaphore from deepfuze.typing import Face, VisionFrame, UpdateProgress, ProcessMode, ModelSet, OptionsWithModel, AudioFrame, QueuePayload from deepfuze.filesystem import is_file, has_audio, resolve_relative_path from deepfuze.download import conditional_download, is_download_done from deepfuze.audio import read_static_voice, get_voice_frame, create_empty_audio_frame from deepfuze.filesystem import is_image, is_video, filter_audio_paths from deepfuze.common_helper import get_first from deepfuze.vision import read_image, read_static_image, write_image, restrict_video_fps from deepfuze.processors.frame.typings import LipSyncerInputs from deepfuze.voice_extractor import clear_voice_extractor from deepfuze.processors.frame import globals as frame_processors_globals from deepfuze.processors.frame import choices as frame_processors_choices FRAME_PROCESSOR = None NAME = __name__.upper() MODELS : ModelSet =\ { 'wav2lip_gan': { 'url': 'https://github.com/facefusion/facefusion-assets/releases/download/models/wav2lip_gan.onnx', 'path': resolve_relative_path('../../../models/deepfuze/wav2lip_gan.onnx') } } OPTIONS : Optional[OptionsWithModel] = None def get_frame_processor() -> Any: global FRAME_PROCESSOR with thread_lock(): while process_manager.is_checking(): sleep(0.5) if FRAME_PROCESSOR is None: model_path = get_options('model').get('path') FRAME_PROCESSOR = onnxruntime.InferenceSession(model_path, providers = apply_execution_provider_options(deepfuze.globals.execution_device_id, deepfuze.globals.execution_providers)) return FRAME_PROCESSOR def clear_frame_processor() -> None: global FRAME_PROCESSOR FRAME_PROCESSOR = None def get_options(key : Literal['model']) -> Any: global OPTIONS if OPTIONS is None: OPTIONS =\ { 'model': MODELS[frame_processors_globals.lip_syncer_model] } return OPTIONS.get(key) def set_options(key : Literal['model'], value : Any) -> None: global OPTIONS OPTIONS[key] = value def register_args(program : ArgumentParser) -> None: program.add_argument('--lip-syncer-model', help = wording.get('help.lip_syncer_model'), default = config.get_str_value('frame_processors.lip_syncer_model', 'wav2lip_gan'), choices = frame_processors_choices.lip_syncer_models) def apply_args(program : ArgumentParser) -> None: args = program.parse_args() frame_processors_globals.lip_syncer_model = args.lip_syncer_model def pre_check() -> bool: download_directory_path = resolve_relative_path('../../../models/deepfuze') model_url = get_options('model').get('url') model_path = get_options('model').get('path') if not deepfuze.globals.skip_download: process_manager.check() conditional_download(download_directory_path, [ model_url ]) process_manager.end() return is_file(model_path) def post_check() -> bool: model_url = get_options('model').get('url') model_path = get_options('model').get('path') if not deepfuze.globals.skip_download and not is_download_done(model_url, model_path): logger.error(wording.get('model_download_not_done') + wording.get('exclamation_mark'), NAME) return False if not is_file(model_path): logger.error(wording.get('model_file_not_present') + wording.get('exclamation_mark'), NAME) return False return True def pre_process(mode : ProcessMode) -> bool: if not has_audio(deepfuze.globals.source_paths): logger.error(wording.get('select_audio_source') + wording.get('exclamation_mark'), NAME) return False if mode in [ 'output', 'preview' ] and not is_image(deepfuze.globals.target_path) and not is_video(deepfuze.globals.target_path): logger.error(wording.get('select_image_or_video_target') + wording.get('exclamation_mark'), NAME) return False if mode == 'output' and not normalize_output_path(deepfuze.globals.target_path, deepfuze.globals.output_path): logger.error(wording.get('select_file_or_directory_output') + wording.get('exclamation_mark'), NAME) return False return True def post_process() -> None: read_static_image.cache_clear() read_static_voice.cache_clear() if deepfuze.globals.video_memory_strategy == 'strict' or deepfuze.globals.video_memory_strategy == 'moderate': clear_frame_processor() if deepfuze.globals.video_memory_strategy == 'strict': clear_face_analyser() clear_content_analyser() clear_face_occluder() clear_face_parser() clear_voice_extractor() def sync_lip(target_face : Face, temp_audio_frame : AudioFrame, temp_vision_frame : VisionFrame) -> VisionFrame: frame_processor = get_frame_processor() crop_mask_list = [] temp_audio_frame = prepare_audio_frame(temp_audio_frame) crop_vision_frame, affine_matrix = warp_face_by_face_landmark_5(temp_vision_frame, target_face.landmarks.get('5/68'), 'ffhq_512', (512, 512)) face_landmark_68 = cv2.transform(target_face.landmarks.get('68').reshape(1, -1, 2), affine_matrix).reshape(-1, 2) bounding_box = create_bounding_box_from_face_landmark_68(face_landmark_68) bounding_box[1] -= numpy.abs(bounding_box[3] - bounding_box[1]) * 0.125 mouth_mask = create_mouth_mask(face_landmark_68) crop_mask_list.append(mouth_mask) box_mask = create_static_box_mask(crop_vision_frame.shape[:2][::-1], deepfuze.globals.face_mask_blur, deepfuze.globals.face_mask_padding) crop_mask_list.append(box_mask) if 'occlusion' in deepfuze.globals.face_mask_types: occlusion_mask = create_occlusion_mask(crop_vision_frame) crop_mask_list.append(occlusion_mask) close_vision_frame, close_matrix = warp_face_by_bounding_box(crop_vision_frame, bounding_box, (96, 96)) close_vision_frame = prepare_crop_frame(close_vision_frame) with conditional_thread_semaphore(deepfuze.globals.execution_providers): close_vision_frame = frame_processor.run(None, { 'source': temp_audio_frame, 'target': close_vision_frame })[0] crop_vision_frame = normalize_crop_frame(close_vision_frame) crop_vision_frame = cv2.warpAffine(crop_vision_frame, cv2.invertAffineTransform(close_matrix), (512, 512), borderMode = cv2.BORDER_REPLICATE) crop_mask = numpy.minimum.reduce(crop_mask_list) paste_vision_frame = paste_back(temp_vision_frame, crop_vision_frame, crop_mask, affine_matrix) return paste_vision_frame def prepare_audio_frame(temp_audio_frame : AudioFrame) -> AudioFrame: temp_audio_frame = numpy.maximum(numpy.exp(-5 * numpy.log(10)), temp_audio_frame) temp_audio_frame = numpy.log10(temp_audio_frame) * 1.6 + 3.2 temp_audio_frame = temp_audio_frame.clip(-4, 4).astype(numpy.float32) temp_audio_frame = numpy.expand_dims(temp_audio_frame, axis = (0, 1)) return temp_audio_frame def prepare_crop_frame(crop_vision_frame : VisionFrame) -> VisionFrame: crop_vision_frame = numpy.expand_dims(crop_vision_frame, axis = 0) prepare_vision_frame = crop_vision_frame.copy() prepare_vision_frame[:, 48:] = 0 crop_vision_frame = numpy.concatenate((prepare_vision_frame, crop_vision_frame), axis = 3) crop_vision_frame = crop_vision_frame.transpose(0, 3, 1, 2).astype('float32') / 255.0 return crop_vision_frame def normalize_crop_frame(crop_vision_frame : VisionFrame) -> VisionFrame: crop_vision_frame = crop_vision_frame[0].transpose(1, 2, 0) crop_vision_frame = crop_vision_frame.clip(0, 1) * 255 crop_vision_frame = crop_vision_frame.astype(numpy.uint8) return crop_vision_frame def get_reference_frame(source_face : Face, target_face : Face, temp_vision_frame : VisionFrame) -> VisionFrame: pass def process_frame(inputs : LipSyncerInputs) -> VisionFrame: reference_faces = inputs.get('reference_faces') source_audio_frame = inputs.get('source_audio_frame') target_vision_frame = inputs.get('target_vision_frame') if deepfuze.globals.face_selector_mode == 'many': many_faces = get_many_faces(target_vision_frame) if many_faces: for target_face in many_faces: target_vision_frame = sync_lip(target_face, source_audio_frame, target_vision_frame) if deepfuze.globals.face_selector_mode == 'one': target_face = get_one_face(target_vision_frame) if target_face: target_vision_frame = sync_lip(target_face, source_audio_frame, target_vision_frame) if deepfuze.globals.face_selector_mode == 'reference': similar_faces = find_similar_faces(reference_faces, target_vision_frame, deepfuze.globals.reference_face_distance) if similar_faces: for similar_face in similar_faces: target_vision_frame = sync_lip(similar_face, source_audio_frame, target_vision_frame) return target_vision_frame def process_frames(source_paths : List[str], queue_payloads : List[QueuePayload], update_progress : UpdateProgress) -> None: reference_faces = get_reference_faces() if 'reference' in deepfuze.globals.face_selector_mode else None source_audio_path = get_first(filter_audio_paths(source_paths)) temp_video_fps = restrict_video_fps(deepfuze.globals.target_path, deepfuze.globals.output_video_fps) for queue_payload in process_manager.manage(queue_payloads): frame_number = queue_payload['frame_number'] target_vision_path = queue_payload['frame_path'] source_audio_frame = get_voice_frame(source_audio_path, temp_video_fps, frame_number) if not numpy.any(source_audio_frame): source_audio_frame = create_empty_audio_frame() target_vision_frame = read_image(target_vision_path) output_vision_frame = process_frame( { 'reference_faces': reference_faces, 'source_audio_frame': source_audio_frame, 'target_vision_frame': target_vision_frame }) write_image(target_vision_path, output_vision_frame) update_progress(1) def process_image(source_paths : List[str], target_path : str, output_path : str) -> None: reference_faces = get_reference_faces() if 'reference' in deepfuze.globals.face_selector_mode else None source_audio_frame = create_empty_audio_frame() target_vision_frame = read_static_image(target_path) output_vision_frame = process_frame( { 'reference_faces': reference_faces, 'source_audio_frame': source_audio_frame, 'target_vision_frame': target_vision_frame }) write_image(output_path, output_vision_frame) def process_video(source_paths : List[str], temp_frame_paths : List[str]) -> None: source_audio_paths = filter_audio_paths(deepfuze.globals.source_paths) temp_video_fps = restrict_video_fps(deepfuze.globals.target_path, deepfuze.globals.output_video_fps) for source_audio_path in source_audio_paths: read_static_voice(source_audio_path, temp_video_fps) frame_processors.multi_process_frames(source_paths, temp_frame_paths, process_frames)