不会画线稿？PaddlePaddle让你秒变灵魂画手！

不会画线稿？PaddlePaddle让你秒变灵魂画手！

线稿可以让空无一物的画纸产生正形负形，更能以长短虚实、疏密深淡、张弛得当之势自然勾勒物象之形、神、光、色、体积、质感等，不同造诣的画者能驾驭出不同的画面，难度之大深不可测，变化多端甚是神奇。

线描技法源远流长，可以追溯到我们中国画的白描，中国古代有许多白描大师，如顾恺之、李公麟等都为我国留下了文艺瑰宝。

但线稿的另一种意义是从某个图片转变而来，只有黑色线条，便于临摹，十分方便。

PaddleHub转换

由开发者Mr.郑先生_提供
【PaddleHub模型贡献】一行代码实现从彩色图提取素描线稿转换成PaddleHub让大家更方便的使用

一、效果展示

图片效果

视频效果

原始视频链接：
https://player.bilibili.com/player.html?aid=373068021&bvid=BV14Z4y1g7uG&cid=264240737&page=1

将视频内容转换为线稿：
https://player.bilibili.com/player.html?aid=800614383&bvid=BV1zy4y1v7kQ&cid=262900480&page=1

二、实现步骤

1.导入必要的库和模型

import cv2
from scipy import ndimage
from model import Model
import numpy as np
import os
os.environ['CUDA_VISIBLE_DEVICES']='0'
model = Model('inference_model',use_gpu=True,use_mkldnn=False,combined=False)

2.处理视频

将视频按帧进行处理，并保存到images文件夹中。

def transform_video_to_image(video_file_path, img_path):'''将视频中每一帧保存成图片'''video_capture = cv2.VideoCapture(video_file_path)fps = video_capture.get(cv2.CAP_PROP_FPS)count = 0while(True):ret, frame = video_capture.read() if ret:cv2.imwrite(img_path + '%d.jpg' % count, frame)count += 1else:breakvideo_capture.release()print('视频图片保存成功, 共有 %d 张' % count)return fps
fps = transform_video_to_image('shipin.mp4', 'images/')

3.图片线稿化

from function import *for home, dirs, files in os.walk('images'):for filename in files:fullname = os.path.join(home, filename)from_mat = cv2.imread(fullname)width = float(from_mat.shape[1])height = float(from_mat.shape[0])new_width = 0new_height = 0if (width > height):from_mat = cv2.resize(from_mat, (512, int(512 / width * height)), interpolation=cv2.INTER_AREA)new_width = 512new_height = int(512 / width * height)else:from_mat = cv2.resize(from_mat, (int(512 / height * width), 512), interpolation=cv2.INTER_AREA)new_width = int(512 / height * width)new_height = 512from_mat = from_mat.transpose((2, 0, 1))light_map = np.zeros(from_mat.shape, dtype=np.float)for channel in range(3):light_map[channel] = get_light_map_single(from_mat[channel])light_map = normalize_pic(light_map)light_map = resize_img_512_3d(light_map)light_map = light_map.astype('float32')line_mat = model.predict(np.expand_dims(light_map, axis=0).astype('float32'))# 去除 batch 维度 (512, 512, 3)line_mat = line_mat.transpose((3, 1, 2, 0))[0]# 裁剪 (512, 384, 3)line_mat = line_mat[0:int(new_height), 0:int(new_width), :]line_mat = np.amax(line_mat, 2)# 降噪show_active_img_and_save_denoise(line_mat, './output/' + filename)print('图片' + filename + '已经完成')
print('全部图片转换成功。')

4.合并视频

def combine_image_to_video(comb_path, output_file_path, fps=30, is_print=False):'''合并图像到视频'''fourcc = cv2.VideoWriter_fourcc(*'MP4V')    file_items = os.listdir(comb_path)file_len = len(file_items)# print(comb_path, file_items)if file_len > 0 :temp_img = cv2.imread(os.path.join(comb_path, file_items[0]))img_height, img_width = temp_img.shape[0], temp_img.shape[1]out = cv2.VideoWriter(output_file_path, fourcc, fps, (img_width, img_height))for i in range(file_len):pic_name = os.path.join(comb_path, str(i)+".jpg")if is_print:print(i+1,'/', file_len, ' ', pic_name)img = cv2.imread(pic_name)out.write(img)out.release()
combine_image_to_video('output', 'work/mp4_analysis.mp4', fps)

5.合并音频

#音频获取
def getMusic(video_name):"""获取指定视频的音频"""# 读取视频文件video = VideoFileClip(video_name)# 返回音频return video.audio
#音频添加
def addMusic(video_name, audio,output_video):"""实现混流，给video_name添加音频"""# 读取视频video = VideoFileClip(video_name)# 设置视频的音频video = video.set_audio(audio)# 保存新的视频文件video.write_videofile(output_video)from moviepy.editor import *
addMusic('work/mp4_analysis.mp4',getMusic('shipin.mp4'),'work/mp4_analysisnew.mp4')

点我进入项目-不会画线稿？PaddlePaddle让你秒变灵魂画手！