本文整理汇总了Python中utils.util.KNNIds方法的典型用法代码示例。如果您正苦于以下问题：Python util.KNNIds方法的具体用法？Python util.KNNIds怎么用？Python util.KNNIds使用的例子？那么, 这里精选的方法代码示例或许可以为您提供帮助。您也可以进一步了解该方法所在类utils.util的用法示例。
在下文中一共展示了util.KNNIds方法的1个代码示例，这些例子默认根据受欢迎程度排序。您可以为喜欢或者感觉有用的代码点赞，您的评价将有助于系统推荐出更棒的Python代码示例。
示例1: GenerateImitationVideo
# 需要导入模块: from utils import util [as 别名]
# 或者: from utils.util import KNNIds [as 别名]
def GenerateImitationVideo(
vid_name, query_ims, query_embs, target_ims, target_embs, height, width):
"""Generates a single cross-sequence imitation video.
For each frame in some query sequence, find the nearest neighbor from
some target sequence in embedding space.
Args:
vid_name: String, the name of the video.
query_ims: Numpy array of shape [query sequence length, height, width, 3].
query_embs: Numpy array of shape [query sequence length, embedding size].
target_ims: Numpy array of shape [target sequence length, height, width,
3].
target_embs: Numpy array of shape [target sequence length, embedding
size].
height: Int, height of the raw image.
width: Int, width of the raw image.
"""
# For each query frame, find the index of the nearest neighbor in the
# target video.
knn_indices = [util.KNNIds(q, target_embs, k=1)[0] for q in query_embs]
# Create and write out the video.
assert knn_indices
knn_ims = np.array([target_ims[k] for k in knn_indices])
MakeImitationVideo(FLAGS.outdir, vid_name, query_ims, knn_ims, height, width)