CV之FD&FA:利用MTCNN的脚本实现对LFW数据集进行FD人脸检测和FA人脸校准

举报
一个处女座的程序猿 发表于 2021/04/02 03:36:52 2021/04/02
【摘要】 CV之FD&FA:利用MTCNN的脚本实现对LFW数据集进行FD人脸检测和FA人脸校准       目录 运行结果 运行过程 运行(部分)代码 在裁剪好的LFW数据集进行验证       运行结果   运行过程 time:14:03~14:37 usag...

CV之FD&FA:利用MTCNN的脚本实现对LFW数据集进行FD人脸检测和FA人脸校准

 

 

 

目录

运行结果

运行过程

运行(部分)代码

在裁剪好的LFW数据集进行验证


 

 

 

运行结果

 

运行过程

time:14:03~14:37


  
  1. usage: git diff [--no-index] <path> <path>
  2. Creating networks and loading parameters
  3. 2018-10-01 14:03:00.211213: I tensorflow/core/platform/cpu_feature_guard.cc:141] Your CPU supports instructions that this TensorFlow binary was not compiled to use: AVX2
  4. WARNING:tensorflow:From F:\AI\DL21TF\DL21examples\chapter_6\src\align\detect_face.py:210: calling reduce_max (from tensorflow.python.ops.math_ops) with keep_dims is deprecated and will be removed in a future version.
  5. Instructions for updating:
  6. keep_dims is deprecated, use keepdims instead
  7. WARNING:tensorflow:From F:\AI\DL21TF\DL21examples\chapter_6\src\align\detect_face.py:212: calling reduce_sum (from tensorflow.python.ops.math_ops) with keep_dims is deprecated and will be removed in a future version.
  8. Instructions for updating:
  9. keep_dims is deprecated, use keepdims instead
  10. datasets/lfw/raw\Shane_Phillips\Shane_Phillips_0001.jpg
  11. F:\Program Files\Python\Python36\Lib\site-packages\scipy\misc\pilutil.py:482: FutureWarning: Conversion of the second argument of issubdtype from `int` to `np.signedinteger` is deprecated. In future, it will be treated as `np.int32 == np.dtype(int).type`.
  12. if issubdtype(ts, int):
  13. F:\Program Files\Python\Python36\Lib\site-packages\scipy\misc\pilutil.py:485: FutureWarning: Conversion of the second argument of issubdtype from `float` to `np.floating` is deprecated. In future, it will be treated as `np.float64 == np.dtype(float).type`.
  14. elif issubdtype(type(size), float):
  15. datasets/lfw/raw\Alejandro_Lopez\Alejandro_Lopez_0001.jpg
  16. ……
  17. datasets/lfw/raw\Recep_Tayyip_Erdogan\Recep_Tayyip_Erdogan_0010.jpg
  18. ……
  19. datasets/lfw/raw\Ruth_Pearce\Ruth_Pearce_0001.jpg
  20. datasets/lfw/raw\Claudio_Abbado\Claudio_Abbado_0001.jpg
  21. datasets/lfw/raw\Dean_Jacek\Dean_Jacek_0001.jpg
  22. datasets/lfw/raw\Mark_Leno\Mark_Leno_0001.jpg
  23. datasets/lfw/raw\Kirk_Franklin\Kirk_Franklin_0001.jpg
  24. datasets/lfw/raw\Doris_Schroeder\Doris_Schroeder_0004.jpg
  25. datasets/lfw/raw\Doris_Schroeder\Doris_Schroeder_0002.jpg
  26. datasets/lfw/raw\Doris_Schroeder\Doris_Schroeder_0001.jpg
  27. datasets/lfw/raw\Doris_Schroeder\Doris_Schroeder_0003.jpg
  28. datasets/lfw/raw\Anne_Heche\Anne_Heche_0001.jpg
  29. datasets/lfw/raw\Keith_Fotta\Keith_Fotta_0001.jpg
  30. datasets/lfw/raw\Zdravko_Mucic\Zdravko_Mucic_0001.jpg
  31. datasets/lfw/raw\Mohammad_Fares\Mohammad_Fares_0001.jpg
  32. datasets/lfw/raw\Bruce_Lunsford\Bruce_Lunsford_0001.jpg
  33. datasets/lfw/raw\Craig_David\Craig_David_0001.jpg
  34. datasets/lfw/raw\Christine_Todd_Whitman\Christine_Todd_Whitman_0003.jpg
  35. datasets/lfw/raw\Christine_Todd_Whitman\Christine_Todd_Whitman_0006.jpg
  36. datasets/lfw/raw\Christine_Todd_Whitman\Christine_Todd_Whitman_0002.jpg
  37. datasets/lfw/raw\Christine_Todd_Whitman\Christine_Todd_Whitman_0001.jpg
  38. datasets/lfw/raw\Christine_Todd_Whitman\Christine_Todd_Whitman_0004.jpg
  39. datasets/lfw/raw\Christine_Todd_Whitman\Christine_Todd_Whitman_0005.jpg
  40. datasets/lfw/raw\Maritza_Macias_Furano\Maritza_Macias_Furano_0001.jpg
  41. datasets/lfw/raw\Edward_Burns\Edward_Burns_0001.jpg
  42. datasets/lfw/raw\Jeff_Roehm\Jeff_Roehm_0001.jpg
  43. datasets/lfw/raw\Mary_Frances_Seiter\Mary_Frances_Seiter_0001.jpg
  44. datasets/lfw/raw\Roger_Moore\Roger_Moore_0002.jpg
  45. datasets/lfw/raw\Roger_Moore\Roger_Moore_0003.jpg
  46. datasets/lfw/raw\Roger_Moore\Roger_Moore_0001.jpg
  47. datasets/lfw/raw\Roger_Moore\Roger_Moore_0005.jpg
  48. datasets/lfw/raw\Roger_Moore\Roger_Moore_0004.jpg
  49. datasets/lfw/raw\Jason_Priestley\Jason_Priestley_0001.jpg
  50. datasets/lfw/raw\Juergen_Chrobog\Juergen_Chrobog_0001.jpg
  51. datasets/lfw/raw\Deb_Santos\Deb_Santos_0001.jpg
  52. datasets/lfw/raw\Javier_Vazquez\Javier_Vazquez_0001.jpg
  53. datasets/lfw/raw\Vanessa_Laine\Vanessa_Laine_0001.jpg
  54. datasets/lfw/raw\Ilan_Goldfajn\Ilan_Goldfajn_0001.jpg
  55. datasets/lfw/raw\Tom_Osborne\Tom_Osborne_0001.jpg
  56. datasets/lfw/raw\Colleen_Donovan\Colleen_Donovan_0001.jpg
  57. datasets/lfw/raw\Jeff_Feldman\Jeff_Feldman_0001.jpg
  58. datasets/lfw/raw\DAngelo_Jimenez\DAngelo_Jimenez_0001.jpg
  59. datasets/lfw/raw\Jane_Riley\Jane_Riley_0001.jpg
  60. datasets/lfw/raw\Frank_Keating\Frank_Keating_0001.jpg
  61. datasets/lfw/raw\Rick_Romley\Rick_Romley_0001.jpg
  62. datasets/lfw/raw\Rick_Romley\Rick_Romley_0003.jpg
  63. datasets/lfw/raw\Rick_Romley\Rick_Romley_0002.jpg
  64. datasets/lfw/raw\Alexandre_Daigle\Alexandre_Daigle_0001.jpg
  65. datasets/lfw/raw\Stephen_Funk\Stephen_Funk_0001.jpg
  66. datasets/lfw/raw\Joe_Nichols\Joe_Nichols_0001.jpg
  67. datasets/lfw/raw\Joe_Nichols\Joe_Nichols_0003.jpg
  68. datasets/lfw/raw\Joe_Nichols\Joe_Nichols_0004.jpg
  69. datasets/lfw/raw\Joe_Nichols\Joe_Nichols_0002.jpg
  70. datasets/lfw/raw\Garth_Drabinsky\Garth_Drabinsky_0001.jpg
  71. datasets/lfw/raw\Pascal_Lamy\Pascal_Lamy_0001.jpg
  72. datasets/lfw/raw\Pascal_Lamy\Pascal_Lamy_0002.jpg
  73. datasets/lfw/raw\Raja_Ramani\Raja_Ramani_0001.jpg
  74. datasets/lfw/raw\David_Zeplowitz\David_Zeplowitz_0001.jpg
  75. datasets/lfw/raw\Kobe_Bryant\Kobe_Bryant_0003.jpg
  76. datasets/lfw/raw\Kobe_Bryant\Kobe_Bryant_0002.jpg
  77. datasets/lfw/raw\Kobe_Bryant\Kobe_Bryant_0001.jpg
  78. ……
  79. datasets/lfw/raw\Renee_Zellweger\Renee_Zellweger_0002.jpg
  80. datasets/lfw/raw\Renee_Zellweger\Renee_Zellweger_0004.jpg
  81. datasets/lfw/raw\Renee_Zellweger\Renee_Zellweger_0015.jpg
  82. datasets/lfw/raw\Jack_Nicholson\Jack_Nicholson_0003.jpg
  83. datasets/lfw/raw\Jack_Nicholson\Jack_Nicholson_0001.jpg
  84. datasets/lfw/raw\Jack_Nicholson\Jack_Nicholson_0002.jpg
  85. datasets/lfw/raw\Edward_Egan\Edward_Egan_0001.jpg
  86. datasets/lfw/raw\Ion_Tiriac\Ion_Tiriac_0001.jpg
  87. datasets/lfw/raw\Jack_Knowlton\Jack_Knowlton_0001.jpg
  88. datasets/lfw/raw\Akbar_Al_Baker\Akbar_Al_Baker_0001.jpg
  89. datasets/lfw/raw\Gholamreza_Aghazadeh\Gholamreza_Aghazadeh_0001.jpg
  90. datasets/lfw/raw\Ramon_Cardenas\Ramon_Cardenas_0001.jpg
  91. datasets/lfw/raw\Ted_Maher\Ted_Maher_0002.jpg
  92. datasets/lfw/raw\Ted_Maher\Ted_Maher_0001.jpg
  93. datasets/lfw/raw\Dionyssis_Georgiadis\Dionyssis_Georgiadis_0001.jpg
  94. datasets/lfw/raw\Gennifer_Flowers\Gennifer_Flowers_0001.jpg
  95. datasets/lfw/raw\Bernard_Giraudeau\Bernard_Giraudeau_0001.jpg
  96. datasets/lfw/raw\Mary_Hill\Mary_Hill_0001.jpg
  97. datasets/lfw/raw\John_Leguizamo\John_Leguizamo_0001.jpg
  98. datasets/lfw/raw\Celia_Cruz\Celia_Cruz_0001.jpg
  99. datasets/lfw/raw\Igor_Ivanov\Igor_Ivanov_0001.jpg
  100. ……
  101. datasets/lfw/raw\Johnny_Benson\Johnny_Benson_0001.jpg
  102. datasets/lfw/raw\Ronaldo_Luis_Nazario_de_Lima\Ronaldo_Luis_Nazario_de_Lima_0004.jpg
  103. datasets/lfw/raw\Ronaldo_Luis_Nazario_de_Lima\Ronaldo_Luis_Nazario_de_Lima_0002.jpg
  104. datasets/lfw/raw\Ronaldo_Luis_Nazario_de_Lima\Ronaldo_Luis_Nazario_de_Lima_0003.jpg
  105. datasets/lfw/raw\Ronaldo_Luis_Nazario_de_Lima\Ronaldo_Luis_Nazario_de_Lima_0001.jpg
  106. datasets/lfw/raw\Carroll_Weimer\Carroll_Weimer_0001.jpg
  107. datasets/lfw/raw\Mark_Butcher\Mark_Butcher_0001.jpg
  108. datasets/lfw/raw\Takenori_Kanzaki\Takenori_Kanzaki_0001.jpg
  109. Total number of images: 13233
  110. Number of successfully aligned images: 13233

运行(部分)代码


  
  1. from __future__ import absolute_import
  2. from __future__ import division
  3. from __future__ import print_function
  4. from scipy import misc
  5. import sys
  6. import os
  7. import argparse
  8. import tensorflow as tf
  9. import facenet
  10. import align.detect_face
  11. import random
  12. from time import sleep
  13. def parse_arguments(argv):
  14. parser = argparse.ArgumentParser()
  15. parser.add_argument('input_dir', type=str, help='Directory with unaligned images.')
  16. parser.add_argument('output_dir', type=str, help='Directory with aligned face thumbnails.')
  17. parser.add_argument('--image_size', type=int,
  18. ……
  19. return parser.parse_args(argv)
  20. if __name__ == '__main__':
  21. main(parse_arguments(sys.argv[1:]))

在裁剪好的LFW数据集进行验证


  
  1. Instructions for updating:
  2. To construct input pipelines, use the `tf.data` module.
  3. Runnning forward pass on LFW images
  4. Accuracy: 0.992+-0.003
  5. Validation rate: 0.97467+-0.01477 @ FAR=0.00133
  6. Area Under Curve (AUC): 1.000
  7. Equal Error Rate (EER): 0.007

 

 

文章来源: yunyaniu.blog.csdn.net,作者:一个处女座的程序猿,版权归原作者所有,如需转载,请联系作者。

原文链接:yunyaniu.blog.csdn.net/article/details/101226350

【版权声明】本文为华为云社区用户转载文章,如果您发现本社区中有涉嫌抄袭的内容,欢迎发送邮件进行举报,并提供相关证据,一经查实,本社区将立刻删除涉嫌侵权内容,举报邮箱: cloudbbs@huaweicloud.com
  • 点赞
  • 收藏
  • 关注作者

评论(0

0/1000
抱歉,系统识别当前为高风险访问,暂不支持该操作

全部回复

上滑加载中

设置昵称

在此一键设置昵称,即可参与社区互动!

*长度不超过10个汉字或20个英文字符,设置后3个月内不可修改。

*长度不超过10个汉字或20个英文字符,设置后3个月内不可修改。