利用COCOAPI计算Yolov3训练出的模型的MAP值，复现ap

 Average Precision  (AP) @[ IoU=0.50:0.95 | area=   all | maxDets=100 ] = 0.314
 Average Precision  (AP) @[ IoU=0.50      | area=   all | maxDets=100 ] = 0.559
 Average Precision  (AP) @[ IoU=0.75      | area=   all | maxDets=100 ] = 0.318
 Average Precision  (AP) @[ IoU=0.50:0.95 | area= small | maxDets=100 ] = 0.142
 Average Precision  (AP) @[ IoU=0.50:0.95 | area=medium | maxDets=100 ] = 0.341
 Average Precision  (AP) @[ IoU=0.50:0.95 | area= large | maxDets=100 ] = 0.464
 Average Recall     (AR) @[ IoU=0.50:0.95 | area=   all | maxDets=  1 ] = 0.278
 Average Recall     (AR) @[ IoU=0.50:0.95 | area=   all | maxDets= 10 ] = 0.420
 Average Recall     (AR) @[ IoU=0.50:0.95 | area=   all | maxDets=100 ] = 0.442
 Average Recall     (AR) @[ IoU=0.50:0.95 | area= small | maxDets=100 ] = 0.240
 Average Recall     (AR) @[ IoU=0.50:0.95 | area=medium | maxDets=100 ] = 0.482
 Average Recall     (AR) @[ IoU=0.50:0.95 | area= large | maxDets=100 ] = 0.611

以上是我利用官方给出的yolov3.weights通过计算5k.txt里图片计算出的ap值，已经很贴近论文里作者给出的分辨率为416x416的前提下的31%的值。计算步骤如下：

1、确定测试集：https://pjreddie.com/media/files/coco/5k.part内部含有val2014的5000张图片

2、通过darknet自带的valid命令计算生成被比较的json文件

3、./darknet detector valid cfg/coco.data cfg/yolov3.cfg backup/yolov3.weight注意运行此命令的时候coco.data里的必须添加eval = coco不然生成的文件是class类个txt文件，结果会自动生成在result文件夹下。

4、利用以下python脚本计算ap值

#-*- coding:utf-8 -*-
import matplotlib.pyplot as plt 
from pycocotools.coco import COCO 
from pycocotools.cocoeval import COCOeval 
import numpy as np 
import skimage.io as io 
import pylab,json 
pylab.rcParams['figure.figsize'] = (10.0, 8.0) 
def get_img_id(file_name): 
    ls = [] 
    myset = [] 
    annos = json.load(open(file_name, 'r')) 
    for anno in annos: 
      ls.append(anno['image_id']) 
    myset = {}.fromkeys(ls).keys() 
    return myset 
if __name__ == '__main__': 
    annType = ['segm', 'bbox', 'keypoints']#set iouType to 'segm', 'bbox' or 'keypoints'
    annType = annType[1] # specify type here
    cocoGt_file = '/home/ss/data/coco2014/annotations/instances_val2014.json'
    cocoGt = COCO(cocoGt_file)#取得标注集中coco json对象
    cocoDt_file = '/home/ss/darknet/results/coco_results.json'
    imgIds = get_img_id(cocoDt_file) 
    print len(imgIds)
    cocoDt = cocoGt.loadRes(cocoDt_file)#取得结果集中image json对象
    imgIds = sorted(imgIds)#按顺序排列coco标注集image_id
    imgIds = imgIds[0:5000]#标注集中的image数据
    cocoEval = COCOeval(cocoGt, cocoDt, annType) 
    cocoEval.params.imgIds = imgIds#参数设置
    cocoEval.evaluate()#评价
    cocoEval.accumulate()#积累
    cocoEval.summarize()#总结

改代码添加的路径信息需要根据自己的实际路径添加，计算出来的信息如顶部。

输出信息的具体含义可以参考链接：

https://blog.csdn.net/u014734886/article/details/78831884

https://blog.csdn.net/u014734886/article/details/78831382