NumPy
东狱邪神
这个作者很懒,什么都没留下…
展开
-
计算交并比
边界框的面积交并比 import numpy as np def compute_iou(boxes1, boxes2): boxes1, boxes2 = np.array(boxes1), np.array(boxes2) lt = np.maximum(boxes1[:, None, :2], boxes2[:, :2]) rb = np.minimum(boxes1[:, None, 2:], boxes2[:, 2:]) wh = np.maximum(rb原创 2021-12-16 21:02:13 · 286 阅读 · 0 评论 -
计算欧式距离
利用Numpy计算: import numpy as np def euclidean_distance(row_points): distances = np.linalg.norm(row_points, ord=2, axis=1) return distances if __name__ == '__main__': row_points = np.array([[3, 4]]).repeat(10, axis=0) distances = euclide原创 2021-12-04 11:08:23 · 653 阅读 · 0 评论 -
NumPy每行/每列只保留一个最大值/最小值,其他元素置为0
import numpy as np input = np.random.randint(100, size=(3, 4)) indices = np.argmax(input, axis=1) indices = np.expand_dims(indices, axis=1) output = np.zeros_like(input) np.put_along_axis(output, indices, 1, axis=1)原创 2021-06-15 11:49:00 · 1948 阅读 · 2 评论