只输入图片在Deformable-ConvNets工程上得出检测结果

我是将数据转化为VOC格式，所以修改Deformable-ConvNets\lib\dataset\pascal_voc.py文件：

注释掉load_pascal_annotation函数中

filename = os.path.join(self.data_path, 'Annotations', index + '.xml')
        tree = ET.parse(filename)
        size = tree.find('size')
        roi_rec['height'] = float(size.find('height').text)
        roi_rec['width'] = float(size.find('width').text)
        #im_size = cv2.imread(roi_rec['image'], cv2.IMREAD_COLOR|cv2.IMREAD_IGNORE_ORIENTATION).shape
        #assert im_size[0] == roi_rec['height'] and im_size[1] == roi_rec['width']

        objs = tree.findall('object')
        if not self.config['use_diff']:
            non_diff_objs = [obj for obj in objs if int(obj.find('difficult').text) == 0]
            objs = non_diff_objs



将
num_objs = len(objs)

改为

num_objs = 1

注释掉：

        for ix, obj in enumerate(objs):
            bbox = obj.find('bndbox')
            # Make pixel indexes 0-based
            x1 = float(bbox.find('xmin').text)
            y1 = float(bbox.find('ymin').text)
            x2 = float(bbox.find('xmax').text)
            y2 = float(bbox.find('ymax').text)
            cls = class_to_index[obj.find('name').text.lower().strip()]
            boxes[ix, :] = [x1, y1, x2, y2]
            gt_classes[ix] = cls
            overlaps[ix, cls] = 1.0
注释掉evaluate_detections中的：

info = self.do_python_eval()
return info

然后将tester.py中pred_eval函数中

info_str = imdb.evaluate_detections(all_boxes)
        if logger:
            logger.info('evaluate detections: \n{}'.format(info_str))

改为：

imdb.evaluate_detections(all_boxes)