【目标检测】YOLOv2--k_means方法

计算公式：

已知：$\mathbf{b}\mathbf{o}{\mathbf{x}}_1=(w_1,h_1)$，$\mathbf{b}\mathbf{o}{\mathbf{x}}_2=(w_2,h_2)$

$\mathbf{b}\mathbf{o}{\mathbf{x}}_1$ 和 $\mathbf{b}\mathbf{o}{\mathbf{x}}_2$ 的交并比：

$$J({\mathbf{b}}_1,{\mathbf{b}}_2)=\frac{\text{min}(w_1,w_2)\cdot \text{min}(h_1,h_2)}{w_1{h}_1+w_2{h}_2-\text{min}(w_1,w_2)\cdot \text{min}(h_1,h_2)}$$

YOLOv2中不使用欧式距离，使用如下公式来计算两个框的距离

$$d_J({\mathbf{b}}_1,{\mathbf{b}}_2)=1-J({\mathbf{b}}_1,{\mathbf{b}}_2)$$

python实现

import numpy as np

# 计算质心和其他boxes的距离
def iou(box, clusters):
    x = np.minimum(clusters[:, 0], box[0])
    y = np.minimum(clusters[:, 1], box[1])

    intersection = x * y
    box_area = box[0] * box[1]
    cluster_area = clusters[:, 0] * clusters[:, 1]

    iou_ = intersection / (box_area + cluster_area - intersection)

    return iou_

def kmeans(boxes, k, dist=np.median):
    rows = boxes.shape[0]

    # rows行k列，每行对应一个boxes，每列是boxes与各个质心的距离
    distances = np.empty((rows, k))
    #
    last_clusters = np.zeros((rows,))

    np.random.seed()
    # 在boxes中随机选取k个作为质心clusters
    clusters = boxes[np.random.choice(rows, k, replace=False)]

    while True:
        for row in range(rows):
            distances[row] = 1 - iou(boxes[row], clusters)
        # 找到与boxes距离最近的clusters的索引
        nearest_clusters = np.argmin(distances, axis=1)

        # 当两次聚类结果相同时结束
        if (last_clusters == nearest_clusters).all():
            break

        # 分别取w和h的平均值更新clusters
        for cluster in range(k):
            clusters[cluster] = dist(boxes[nearest_clusters == cluster], axis=0)

        last_clusters = nearest_clusters

    return clusters

参考：
https://lars76.github.io/object-detection/k-means-anchor-boxes/#