自定义博客皮肤VIP专享

*博客头图:

格式为PNG、JPG,宽度*高度大于1920*100像素,不超过2MB,主视觉建议放在右侧,请参照线上博客头图

请上传大于1920*100像素的图片!

博客底图:

图片格式为PNG、JPG,不超过1MB,可上下左右平铺至整个背景

栏目图:

图片格式为PNG、JPG,图片宽度*高度为300*38像素,不超过0.5MB

主标题颜色:

RGB颜色,例如:#AFAFAF

Hover:

RGB颜色,例如:#AFAFAF

副标题颜色:

RGB颜色,例如:#AFAFAF

自定义博客皮肤

-+

abc8350712的博客

深度学习

  • 博客(3)
  • 收藏
  • 关注

原创 seglink代码之 generate_anchors_one layer

''' 特别关注np.mgrid函数,例如 x, y = np.mgrid[0:4, 0:4] x = 0, 1, 2, 3 0, 1, 2, 3 0, 1, 2, 3 0, 1, 2, 3y = 0, 0, 0, 0 1, 1, 1, 1 2, 2, 2, 2 3, 3, 3, 3 利用这个函数的性质就能很方便的计算anchor_bo

2017-10-27 19:46:04 1018

原创 Seglink代码之二 数据增强 distorted_bounding_box_crop

def distorted_bounding_box_crop(image, labels, bboxes, xs, ys, min_objec

2017-10-27 19:39:30 1087

原创 Seglink代码之ICDAR2015_to_tfrecords

关键性代码在于 convert_to_example。比较重要的数据有: 1)GT的八个点 2)GT的四个点 3)imagedef convert_to_example(image_data, filename, labels, ignored, labels_text, bboxes, oriented_bboxes, shape): """Build an Example prot

2017-10-24 16:30:34 1330

空空如也

空空如也

TA创建的收藏夹 TA关注的收藏夹

TA关注的人

提示
确定要删除当前文章?
取消 删除