前往小程序,Get更优阅读体验!
立即前往
首页
学习
活动
专区
工具
TVP
发布
社区首页 >专栏 >PASCAL VOC的评估代码voc_eval.py解析

PASCAL VOC的评估代码voc_eval.py解析

作者头像
狼啸风云
修改2022-09-03 19:45:18
1.7K0
修改2022-09-03 19:45:18
举报
文章被收录于专栏:计算机视觉理论及其实现

目录

1、读检测的结果

2、解析一幅图像中的目标数

3、计算AP

4、VOC的评估

5、进行python评估

6、voc的检测评估


1、读检测的结果

代码语言:javascript
复制
def write_voc_results_file(all_boxes, test_imgid_list, det_save_dir):
  for cls, cls_id in NAME_LABEL_MAP.items():
    if cls == 'back_ground':
      continue
    print("Writing {} VOC resutls file".format(cls))

    mkdir(det_save_dir)
    det_save_path = os.path.join(det_save_dir, "det_"+cls+".txt")
    with open(det_save_path, 'wt') as f:
      for index, img_name in enumerate(test_imgid_list):
        this_img_detections = all_boxes[index]

        this_cls_detections = this_img_detections[this_img_detections[:, 0]==cls_id]
        if this_cls_detections.shape[0] == 0:
          continue # this cls has none detections in this img
        for a_det in this_cls_detections:
          f.write('{:s} {:.3f} {:.1f} {:.1f} {:.1f} {:.1f}\n'.
                  format(img_name, a_det[1],
                         a_det[2], a_det[3],
                         a_det[4], a_det[5]))  # that is [img_name, score, xmin, ymin, xmax, ymax]

参数:

  • all_boxes:一个列表,每个部件代表一幅图像的检测,检测的结果是一个数组,形状为[-1,6],形式为[category, score, xmin, ymin, xmax, ymax],如果检测到的结果不在图像中,检测是空数组。
  • test_imgid_list:测试的图像列表。
  • det_save_path:检测保存的地址。

2、解析一幅图像中的目标数

代码语言:javascript
复制
def parse_rec(filename):
  """ Parse a PASCAL VOC xml file """
  tree = ET.parse(filename)
  objects = []
  for obj in tree.findall('object'):
    obj_struct = {}
    obj_struct['name'] = obj.find('name').text
    obj_struct['pose'] = obj.find('pose').text
    obj_struct['truncated'] = int(obj.find('truncated').text)
    obj_struct['difficult'] = int(obj.find('difficult').text)
    bbox = obj.find('bndbox')
    obj_struct['bbox'] = [int(bbox.find('xmin').text),
                          int(bbox.find('ymin').text),
                          int(bbox.find('xmax').text),
                          int(bbox.find('ymax').text)]
    objects.append(obj_struct)
  return objects

因为PASCAL VOC的标记格式是xml,此函数作用主要是解析xml文件。

3、计算AP

代码语言:javascript
复制
def voc_ap(rec, prec, use_07_metric=False):
  if use_07_metric:
    # 11 point metric
    ap = 0.
    for t in np.arange(0., 1.1, 0.1):
      if np.sum(rec >= t) == 0:
        p = 0
      else:
        p = np.max(prec[rec >= t])
      ap = ap + p / 11.
  else:
    # correct AP calculation
    # first append sentinel values at the end
    mrec = np.concatenate(([0.], rec, [1.]))
    mpre = np.concatenate(([0.], prec, [0.]))

    # compute the precision envelope
    for i in range(mpre.size - 1, 0, -1):
      mpre[i - 1] = np.maximum(mpre[i - 1], mpre[i])

    # to calculate area under PR curve, look for points
    # where X axis (recall) changes value
    i = np.where(mrec[1:] != mrec[:-1])[0]

    # and sum (\Delta recall) * prec
    ap = np.sum((mrec[i + 1] - mrec[i]) * mpre[i + 1])
  return ap

给定精度和召回率计算VOC的AP,如果use_07_metric为真,使用VOC 07 11点方法。

4、VOC的评估

代码语言:javascript
复制
def voc_eval(detpath, annopath, test_imgid_list, cls_name, ovthresh=0.5,
                 use_07_metric=False, use_diff=False):
  # 1. parse xml to get gtboxes
  # read list of images
  imagenames = test_imgid_list

  recs = {}
  for i, imagename in enumerate(imagenames):
    recs[imagename] = parse_rec(os.path.join(annopath, imagename+'.xml'))
    # if i % 100 == 0:
    #   print('Reading annotation for {:d}/{:d}'.format(
    #     i + 1, len(imagenames)))

  # 2. get gtboxes for this class.
  class_recs = {}
  num_pos = 0
  # if cls_name == 'person':
  #   print ("aaa")
  for imagename in imagenames:
    R = [obj for obj in recs[imagename] if obj['name'] == cls_name]
    bbox = np.array([x['bbox'] for x in R])
    if use_diff:
      difficult = np.array([False for x in R]).astype(np.bool)
    else:
      difficult = np.array([x['difficult'] for x in R]).astype(np.bool)
    det = [False] * len(R)
    num_pos = num_pos + sum(~difficult)  # ignored the diffcult boxes
    class_recs[imagename] = {'bbox': bbox,
                             'difficult': difficult,
                             'det': det} # det means that gtboxes has already been detected

  # 3. read the detection file
  detfile = os.path.join(detpath, "det_"+cls_name+".txt")
  with open(detfile, 'r') as f:
    lines = f.readlines()
  # for a line. that is [img_name, confidence, xmin, ymin, xmax, ymax]
  splitlines = [x.strip().split(' ') for x in lines]  # a list that include a list
  image_ids = [x[0] for x in splitlines]  # img_id is img_name
  confidence = np.array([float(x[1]) for x in splitlines])
  BB = np.array([[float(z) for z in x[2:]] for x in splitlines])
  nd = len(image_ids) # num of detections. That, a line is a det_box.
  tp = np.zeros(nd)
  fp = np.zeros(nd)

  if BB.shape[0] > 0:
    # sort by confidence
    sorted_ind = np.argsort(-confidence)
    sorted_scores = np.sort(-confidence)
    BB = BB[sorted_ind, :]
    image_ids = [image_ids[x] for x in sorted_ind]  #reorder the img_name

    # go down dets and mark TPs and FPs
    for d in range(nd):
      R = class_recs[image_ids[d]]  # img_id is img_name
      bb = BB[d, :].astype(float)
      ovmax = -np.inf
      BBGT = R['bbox'].astype(float)

      if BBGT.size > 0:
        # compute overlaps
        # intersection
        ixmin = np.maximum(BBGT[:, 0], bb[0])
        iymin = np.maximum(BBGT[:, 1], bb[1])
        ixmax = np.minimum(BBGT[:, 2], bb[2])
        iymax = np.minimum(BBGT[:, 3], bb[3])
        iw = np.maximum(ixmax - ixmin + 1., 0.)
        ih = np.maximum(iymax - iymin + 1., 0.)
        inters = iw * ih

        # union
        uni = ((bb[2] - bb[0] + 1.) * (bb[3] - bb[1] + 1.) +
               (BBGT[:, 2] - BBGT[:, 0] + 1.) *
               (BBGT[:, 3] - BBGT[:, 1] + 1.) - inters)

        overlaps = inters / uni
        ovmax = np.max(overlaps)
        jmax = np.argmax(overlaps)

      if ovmax > ovthresh:
        if not R['difficult'][jmax]:
          if not R['det'][jmax]:
            tp[d] = 1.
            R['det'][jmax] = 1
          else:
            fp[d] = 1.
      else:
        fp[d] = 1.

  # 4. get recall, precison and AP
  fp = np.cumsum(fp)
  tp = np.cumsum(tp)
  rec = tp / float(num_pos)
  # avoid divide by zero in case the first detection matches a difficult
  # ground truth
  prec = tp / np.maximum(tp + fp, np.finfo(np.float64).eps)
  ap = voc_ap(rec, prec, use_07_metric)

  return rec, prec, ap

参数:

  • param detpath.
  • param annopath.
  • param test_imgid_list: it 's a list that contains the img_name of test_imgs.
  • param cls_name.
  • param ovthresh.
  • param use_07_metric.
  • param use_diff.

5、进行python评估

代码语言:javascript
复制
def do_python_eval(test_imgid_list, test_annotation_path):
  AP_list = []
  # import matplotlib.pyplot as plt
  # import matplotlib.colors as colors
  # color_list = colors.cnames.keys()[::6]

  for cls, index in NAME_LABEL_MAP.items():
    if cls == 'back_ground':
      continue
    recall, precision, AP = voc_eval(detpath=os.path.join(cfgs.EVALUATE_DIR, cfgs.VERSION),
                                     test_imgid_list=test_imgid_list,
                                     cls_name=cls,
                                     annopath=test_annotation_path,
                                     use_07_metric=cfgs.USE_07_METRIC)
    AP_list += [AP]
    print("cls : {}|| Recall: {} || Precison: {}|| AP: {}".format(cls, recall[-1], precision[-1], AP))
    # plt.plot(recall, precision, label=cls, color=color_list[index])
    # plt.legend(loc='upper right')
    # print(10*"__")
  # plt.show()
  # plt.savefig(cfgs.VERSION+'.jpg')
  print("mAP is : {}".format(np.mean(AP_list)))

6、voc的检测评估

代码语言:javascript
复制
def voc_evaluate_detections(all_boxes, test_annotation_path, test_imgid_list):
  '''
  :param all_boxes: is a list. each item reprensent the detections of a img.The detections is a array. shape is [-1, 6]. [category, score, xmin, ymin, xmax, ymax].Note that: if none detections in this img. that the detetions is : []
  :return:
  '''
  test_imgid_list = [item.split('.')[0] for item in test_imgid_list]
  write_voc_results_file(all_boxes, test_imgid_list=test_imgid_list,
                         det_save_dir=os.path.join(cfgs.EVALUATE_DIR, cfgs.VERSION))
  do_python_eval(test_imgid_list, test_annotation_path=test_annotation_path)

参数:

  • 一个列表,每个部件代表检测到的一幅图像,检测结果是一个数组。
本文参与 腾讯云自媒体同步曝光计划,分享自作者个人站点/博客。
原始发表:2019/11/29 ,如有侵权请联系 cloudcommunity@tencent.com 删除

本文分享自 作者个人站点/博客 前往查看

如有侵权,请联系 cloudcommunity@tencent.com 删除。

本文参与 腾讯云自媒体同步曝光计划  ,欢迎热爱写作的你一起参与!

评论
登录后参与评论
0 条评论
热度
最新
推荐阅读
目录
  • 目录
  • 1、读检测的结果
  • 2、解析一幅图像中的目标数
  • 3、计算AP
  • 4、VOC的评估
  • 5、进行python评估
  • 6、voc的检测评估
领券
问题归档专栏文章快讯文章归档关键词归档开发者手册归档开发者手册 Section 归档