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
| import numpy as np
class SeeingMachinesImpairmentDetector:
"""Seeing Machines损伤检测器"""
def __init__(self):
self.dms_features = DMSFeatureExtractor()
self.fusion_module = MultiModalFusion()
def detect(self, camera_data, vehicle_data=None,
steering_data=None) -> dict:
"""
多模态损伤检测
Args:
camera_data: 摄像头数据
vehicle_data: 车辆动力学数据（可选）
steering_data: 方向盘数据（可选）
Returns:
impairment_score: 损伤评分
"""
dms_features = self.dms_features.extract(camera_data)
if vehicle_data and steering_data:
fused_score = self.fusion_module.fuse(
dms_features, vehicle_data, steering_data
)
else:
fused_score = self.dms_features.classify(dms_features)
return {
'impairment_score': fused_score,
'confidence': self._calc_confidence(fused_score),
'modality': 'multi-modal' if vehicle_data else 'vision-only'
}
def _calc_confidence(self, score):
return min(abs(score - 0.5) * 2, 1.0)
class DMSFeatureExtractor:
def extract(self, data):
return {
'gaze_entropy': np.random.rand(),
'blink_pattern': np.random.rand(),
'pupil_response': np.random.rand(),
'facial_rigidity': np.random.rand(),
}
def classify(self, features):
weights = {'gaze_entropy': 0.3, 'blink_pattern': 0.25,
'pupil_response': 0.25, 'facial_rigidity': 0.2}
return sum(w * features[k] for k, w in weights.items())
class MultiModalFusion:
def fuse(self, dms, vehicle, steering):
dms_score = sum(dms.values()) / len(dms)
return 0.6 * dms_score + 0.2 * vehicle + 0.2 * steering
|