gfpgan_enhancer.py

import os
import cv2
import torch
from gfpgan import GFPGANer
from tqdm import tqdm
from basicsr.archs.rrdbnet_arch import RRDBNet
from realesrgan import RealESRGANer

def load_video_to_cv2(input_path):
    video_stream = cv2.VideoCapture(input_path)
    fps = video_stream.get(cv2.CAP_PROP_FPS)
    full_frames = []
    while True:
        still_reading, frame = video_stream.read()
        if not still_reading:
            video_stream.release()
            break
        full_frames.append(cv2.cvtColor(frame, cv2.COLOR_BGR2RGB))
    return full_frames, fps

def save_frames_to_video(frames, output_path, fps):
    if len(frames) == 0:
        raise ValueError("No frames to write to video.")
    
    height, width, _ = frames[0].shape
    fourcc = cv2.VideoWriter_fourcc(*'mp4v')
    video_writer = cv2.VideoWriter(output_path, fourcc, fps, (width, height))
    
    for frame in frames:
        video_writer.write(cv2.cvtColor(frame, cv2.COLOR_RGB2BGR))
    
    video_writer.release()

def process_video_with_gfpgan(input_video_path, output_video_path, model_path='gfpgan/weights/GFPGANv1.4.pth'):
    # Load video and convert to frames
    frames, fps = load_video_to_cv2(input_video_path)


    realesrgan_model = RRDBNet(num_in_ch=3, num_out_ch=3, num_feat=64, num_block=23, num_grow_ch=32, scale=2)
    bg_upsampler = RealESRGANer(
        scale=2,
        model_path="gfpgan/weights/RealESRGAN_x2plus.pth",
        model=realesrgan_model,
        tile=400,
        tile_pad=10,
        pre_pad=0,
        half=True)
    
    # Set up GFPGAN restorer
    arch = 'clean'
    channel_multiplier = 2
    restorer = GFPGANer(
        model_path=model_path,
        upscale=2,
        arch=arch,
        channel_multiplier=channel_multiplier,
        bg_upsampler=bg_upsampler
    )

    # Enhance each frame
    enhanced_frames = []
    print("Enhancing frames...")
    for frame in tqdm(frames, desc='Processing Frames'):
        # Enhance face in the frame
        img = cv2.cvtColor(frame, cv2.COLOR_RGB2BGR)
        _, _, enhanced_img = restorer.enhance(img, has_aligned=False, only_center_face=False, paste_back=True)
        enhanced_frames.append(cv2.cvtColor(enhanced_img, cv2.COLOR_BGR2RGB))

    # Save the enhanced frames to a video
    save_frames_to_video(enhanced_frames, output_video_path, fps)
    print(f'Enhanced video saved at {output_video_path}')