-
Notifications
You must be signed in to change notification settings - Fork 158
/
eval_coco.py
161 lines (140 loc) · 6 KB
/
eval_coco.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
from ctypes import c_float,c_int,Structure,POINTER,CDLL,RTLD_GLOBAL,c_char_p,c_void_p,c_long
import cv2
import numpy as np
import os
from pycocotools.cocoeval import COCOeval
import pycocotools.coco as coco
from tqdm import tqdm
import sys
class BOX(Structure):
_fields_ = [("x1", c_float),
("y1", c_float),
("x2", c_float),
("y2", c_float)]
class landmarks(Structure):
_fields_ = [("x", c_float),
("y", c_float)]
class DETECTION(Structure):
_fields_ = [("bbox", BOX),
("classId", c_int),
("prob", c_float),
("marks", landmarks * 5)]
class detRESULT(Structure):
_fields_ = [("num", c_int),
("det", POINTER(DETECTION))]
lib=CDLL("lib/libctdet.so",RTLD_GLOBAL)
init_net=lib.initNet
init_net.argtypes = [c_char_p]
init_net.restype = c_void_p
free=lib.free
free.argtypes = [c_void_p]
free_result=lib.freeResult
free_result.argtypes = [POINTER(detRESULT)]
free_net=lib.freeNet
free_net.argtypes = [c_void_p]
inference=lib.predict
inference.argtypes=[c_void_p,c_void_p,c_int,c_int]
inference.restype=detRESULT
ndarray_image = lib.ndarrayToImage
ndarray_image.argtypes = [POINTER(c_float), POINTER(c_long), POINTER(c_long)]
ndarray_image.restype = c_void_p
set_device = lib.setDevice
set_device.argtypes = [c_int]
def ndarray_to_image(img):
img = img.astype(np.float32)
data = img.ctypes.data_as(POINTER(c_float))
image_p = ndarray_image(data, img.ctypes.shape, img.ctypes.strides)
return image_p
def img_preprocess(image, target_shape=(512,512),mean = (0.485,0.456,0.406),std =(0.229,0.224,0.225)):
mean = np.array(mean,dtype=np.float32).reshape([1,1,3])
std = np.array(std,dtype=np.float32).reshape([1,1,3])
h_target, w_target = target_shape
h_org, w_org, _ = image.shape
resize_ratio = min(1.0 * w_target / w_org, 1.0 * h_target / h_org)
resize_w = int(resize_ratio * w_org)
resize_h = int(resize_ratio * h_org)
image_resized = cv2.resize(image, (resize_w, resize_h))
image_paded = np.full((h_target, w_target, 3), 0).astype(np.uint8)
dw = int((w_target - resize_w) / 2)
dh = int((h_target - resize_h) / 2)
image_paded[dh:resize_h+dh, dw:resize_w+dw,:] = image_resized
image_paded = (image_paded.astype(np.float32)/255. - mean) / std
return image_paded
def show_img(pred,image):
image_h, image_w, _ = image.shape
bbox_thick = int(1.0 * (image_h + image_w) / 600)
for j in range(pred.num):
x1, y1, x2, y2, classId, prob = pred.det[j].bbox.x1, pred.det[j].bbox.y1, \
pred.det[j].bbox.x2, pred.det[j].bbox.y2, pred.det[j].classId, pred.det[j].prob
cls=int(classId)
cv2.rectangle(image, (int(x1), int(y1)), (int(x2), int(y2)),(255,0,0), bbox_thick)
cv2.putText(image, '%s %f'%(class_name[cls],prob), (int(x2), int(y2)), 0, 0.001 * image_h, (0, 255, 0),
bbox_thick // 2)
return image
def top_k(pred,K=100):
boxs_list = []
for j in range(pred.num):
x1, y1, x2, y2, classId, prob = pred.det[j].bbox.x1, pred.det[j].bbox.y1, \
pred.det[j].bbox.x2, pred.det[j].bbox.y2, pred.det[j].classId, pred.det[j].prob
boxs_list.append([x1,y1,x2,y2,classId,prob])
boxs_list = np.array(boxs_list)
if len(boxs_list)>K:
prob = boxs_list[:,-1]
top_arg = np.argsort(prob)[::-1][:K]
boxs_list = boxs_list[top_arg]
return boxs_list
### you need config include/ctdetConfig.h classNum and visThresh
## for coco
class_name = [
'person', 'bicycle', 'car', 'motorcycle', 'airplane',
'bus', 'train', 'truck', 'boat', 'traffic light', 'fire hydrant',
'stop sign', 'parking meter', 'bench', 'bird', 'cat', 'dog', 'horse',
'sheep', 'cow', 'elephant', 'bear', 'zebra', 'giraffe', 'backpack',
'umbrella', 'handbag', 'tie', 'suitcase', 'frisbee', 'skis',
'snowboard', 'sports ball', 'kite', 'baseball bat', 'baseball glove',
'skateboard', 'surfboard', 'tennis racket', 'bottle', 'wine glass',
'cup', 'fork', 'knife', 'spoon', 'bowl', 'banana', 'apple', 'sandwich',
'orange', 'broccoli', 'carrot', 'hot dog', 'pizza', 'donut', 'cake',
'chair', 'couch', 'potted plant', 'bed', 'dining table', 'toilet', 'tv',
'laptop', 'mouse', 'remote', 'keyboard', 'cell phone', 'microwave',
'oven', 'toaster', 'sink', 'refrigerator', 'book', 'clock', 'vase',
'scissors', 'teddy bear', 'hair drier', 'toothbrush']
valid_ids = [
1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 13,
14, 15, 16, 17, 18, 19, 20, 21, 22, 23,
24, 25, 27, 28, 31, 32, 33, 34, 35, 36,
37, 38, 39, 40, 41, 42, 43, 44, 46, 47,
48, 49, 50, 51, 52, 53, 54, 55, 56, 57,
58, 59, 60, 61, 62, 63, 64, 65, 67, 70,
72, 73, 74, 75, 76, 77, 78, 79, 80, 81,
82, 84, 85, 86, 87, 88, 89, 90]
coco_val_ann = '/data/DataSet/coco2017/annotations/instances_val2017.json'
coco_val_dir = '/data/DataSet/coco2017/val2017'
data = coco.COCO(coco_val_ann)
set_device(0)
net = init_net(bytes(sys.argv[1],encoding = "utf8"))
detections = []
for img_id in tqdm(data.getImgIds()):
img_name = os.path.join(coco_val_dir,data.loadImgs(ids=[img_id])[0]['file_name']).strip()
frame = cv2.imread(img_name)
img_h, img_w, _ = frame.shape
img = img_preprocess(frame, (512, 512), mean=(0.408, 0.447, 0.470), std=(0.289, 0.274, 0.278))
img_p = ndarray_to_image(img)
pred = inference(net,img_p,img_w,img_h)
boxs = top_k(pred)
for i,det in enumerate(boxs):
x, y, x1, y1, cls, conf = det[:6]
detection = {
"image_id": img_id,
"category_id": int(valid_ids[int(cls)]),
"bbox": [x, y, x1 - x, y1 - y],
"score": float("{:.2f}".format(conf))
}
detections.append(detection)
free_result(pred)
free(img_p)
coco_dets = data.loadRes(detections)
coco_eval = COCOeval(data, coco_dets, "bbox")
coco_eval.evaluate()
coco_eval.accumulate()
coco_eval.summarize()