代码拉取完成,页面将自动刷新
同步操作将从 yinguobing/face-mesh-generator 强制同步,此操作会覆盖自 Fork 仓库以来所做的任何修改,且无法恢复!!!
确定后同步将在后台操作,完成时将刷新页面,请耐心等待。
import cv2
import numpy as np
import tensorflow as tf
from mesh_record_operator import MeshRecordOperator
from fmd.mark_dataset.util import draw_marks
if __name__ == "__main__":
dataset = MeshRecordOperator(
"/home/robin/Developer/face-mesh-generator/tfrecord/300w.record")
for sample in dataset.parse_dataset():
image_decoded = tf.image.decode_image(sample['image/encoded']).numpy()
height = sample['image/height'].numpy()
width = sample['image/width'].numpy()
depth = sample['image/depth'].numpy()
filename = sample['image/filename'].numpy()
marks = tf.io.parse_tensor(sample['label/mesh'], tf.float32).numpy()
print(filename, width, height, depth)
# Use OpenCV to preview the image.
image = np.array(image_decoded, np.uint8)
image = cv2.cvtColor(image, cv2.COLOR_RGB2BGR)
# Draw the landmark on image
landmark = np.reshape(marks, (-1, 3)) * width
draw_marks(image, landmark, 1)
# Show the result
cv2.imshow("image", cv2.resize(image, (512, 512)))
if cv2.waitKey() == 27:
break
此处可能存在不合适展示的内容,页面不予展示。您可通过相关编辑功能自查并修改。
如您确认内容无涉及 不当用语 / 纯广告导流 / 暴力 / 低俗色情 / 侵权 / 盗版 / 虚假 / 无价值内容或违法国家有关法律法规的内容,可点击提交进行申诉,我们将尽快为您处理。