在test.py中增加可视化结果的代码后，效果很差。

[liuqinglong110]

您好，我在test.py脚本的test_net()函数内部将检测后的bbox输出在图片中，可是现实效果很差，几乎是没有检测效果。但是，我运行test.py时打印结果和readme中差不多。麻烦您看看是否是我添加可视化结果的代码有问题。

以下是打印结果：
Finished loading model!
loading annotations into memory...
Done (t=0.46s)
creating index...
index created!
val2017 gt roidb loaded from /media/sfpc/hdd01/works/01member/liuqinglong/PyTorch/CFENet/data/coco_cache/val2017_gt_roidb.pkl
=> Total 5000 images to test.
Begin to evaluate
100%|██████████████████████████████████████████████████████████████████████████████████████| 5000/5000 [02:48<00:00, 31.19it/s]
Evaluating detections
Collecting Results......
Writing results json to eval/COCO/detections_val2017_results.json
Loading and preparing results...
DONE (t=1.59s)
creating index...
index created!
Running per image evaluation...
useSegm (deprecated) is not None. Running bbox evaluation
Evaluate annotation type bbox
DONE (t=37.24s).
Accumulating evaluation results...
DONE (t=4.10s).

31.1
[40.57291121621083, 23.09799931924321, 27.079142799861117, 35.90445758180694, 59.03516009559062, 57.60104185561858, 62.10953584207819, 28.346934670516937, 15.879438226694829, 12.970908505667131, 55.03553105170638, 52.27166928240076, 32.502803148994516, 17.25144519844492, 22.92283574501527, 60.524507358106575, 55.43074100407569, 48.30292656303936, 39.25437644104135, 38.50729999007073, 53.78198667257775, 66.53508687457537, 57.069485827333544, 55.48621623615131, 8.226423098688626, 28.16240625576513, 6.334530659774646, 18.606222843150782, 25.961159722050674, 47.57286849340589, 15.05620897410914, 20.209121557732768, 24.63198468746786, 24.948788811398522, 16.490687186409396, 26.951079648205422, 39.290302826940234, 26.357816912911865, 33.12220267646468, 18.87029605885143, 18.62972706367258, 25.819311087372977, 17.209735294725846, 6.039470642626763, 5.763638765228907, 31.060914252819543, 15.184937468529409, 12.08085800131153, 30.269102490882066, 22.767590542989772, 17.840366849920855, 13.39895571754999, 25.849177952756648, 41.56957495448163, 35.43302914486948, 30.78429116989319, 18.60925156812012, 39.383414853228544, 18.458530393041997, 39.194152267453966, 23.626842896745135, 54.8987088811993, 47.88809867681634, 50.07943886845173, 45.832888186451086, 13.44602209563323, 41.22519880390803, 25.22350059982525, 51.13318211095752, 30.908740159536514, 25.331203019967546, 28.81247030147889, 44.56079493823082, 6.662690165252809, 38.242017965681576, 23.895685283325612, 24.197714298866003, 36.17901336876259, 0.0, 9.289938403610515]
~~~~ Summary metrics ~~~~
 Average Precision  (AP) @[ IoU=0.50:0.95 | area=   all | maxDets=100 ] = 0.311
 Average Precision  (AP) @[ IoU=0.50      | area=   all | maxDets=100 ] = 0.494
 Average Precision  (AP) @[ IoU=0.75      | area=   all | maxDets=100 ] = 0.330
 Average Precision  (AP) @[ IoU=0.50:0.95 | area= small | maxDets=100 ] = 0.126
 Average Precision  (AP) @[ IoU=0.50:0.95 | area=medium | maxDets=100 ] = 0.337
 Average Precision  (AP) @[ IoU=0.50:0.95 | area= large | maxDets=100 ] = 0.476
 Average Recall     (AR) @[ IoU=0.50:0.95 | area=   all | maxDets=  1 ] = 0.260
 Average Recall     (AR) @[ IoU=0.50:0.95 | area=   all | maxDets= 10 ] = 0.411
 Average Recall     (AR) @[ IoU=0.50:0.95 | area=   all | maxDets=100 ] = 0.435
 Average Recall     (AR) @[ IoU=0.50:0.95 | area= small | maxDets=100 ] = 0.183
 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 ] = 0.644
Wrote COCO eval results to: eval/COCO/detection_results.pkl
Detect time per image: 0.023s
Nms time per image: 0.005s
Total time per image: 0.029s
FPS: 34.986 fps

我在test.py的90行添加如下代码：
        _t['misc'].tic()
        for j in range(1, num_classes):
            inds = np.where(scores[:, j] > thresh)[0]
            if len(inds) == 0:
                all_boxes[j][i] = np.empty([0, 5], dtype=np.float32)
                continue
            c_bboxes = boxes[inds]
            c_scores = scores[inds, j]
            c_dets = np.hstack((c_bboxes, c_scores[:, np.newaxis])).astype(np.float32, copy=False)

            soft_nms = cfg.test_cfg.soft_nms
            keep = nms(c_dets, cfg.test_cfg.iou, force_cpu=soft_nms)
            keep = keep[:cfg.test_cfg.keep_per_class]
            c_dets = c_dets[keep, :]
            all_boxes[j][i] = c_dets
#**********此处开始及以下6行代码为添加的可视化代码*****************************#
            for id in range(len(c_dets)):
                ymin, xmin, ymax, xmax = c_dets[id,0:4]
                (left, right, top, bottom) = (xmin, xmax, ymin, ymax)
                cv2.rectangle(img, (int(left), int(top)), (int(right), int(bottom)), (255,0,0), 2)
            cv2.imshow('img', img)
            cv2.waitKey(1)
#***************************************************************************#
        if max_per_image > 0:
            image_scores = np.hstack([all_boxes[j][i][:, -1] for j in range(1,num_classes)])
            if len(image_scores) > max_per_image:
                image_thresh = np.sort(image_scores)[-max_per_image]
                for j in range(1, num_classes):
                    keep = np.where(all_boxes[j][i][:, -1] >= image_thresh)[0]
                    all_boxes[j][i] = all_boxes[j][i][keep, :]
        nms_time = _t['misc'].toc()