Information
Contact Name: Carl Lemaire
Email: carl.lemaire@usherbrooke.ca
Organization: Université de Sherbrooke
Method's Name: DenseNet
Reference: Gao Huang, Zhuang Liu, Kilian Q Weinberger, Laurens van der Maaten. Densely connected convolutional networks. arXiv preprint arXiv:1608.06993.
Project Website:
Evaluation Metrics

Precision of each category :

$$ Pre_i = \frac{TP_i}{TP_i + FP_i}$$


Articulated Truck Bicycle Bus Car Motorcycle Non-motorized Vehicle Pedestrian Pickup Truck Single Unit Truck Work Van Background
0.9195 0.8801 0.9887 0.9784 0.9378 0.8385 0.9434 0.9259 0.7554 0.9146 0.9973

Recall of each category:

$$ Rec_i = \frac{TP_i}{TP_i + FN_i}$$


Articulated Truck Bicycle Bus Car Motorcycle Non-motorized Vehicle Pedestrian Pickup Truck Single Unit Truck Work Van Background
0.9142 0.8354 0.9523 0.9857 0.8525 0.5571 0.9585 0.9323 0.7984 0.7919 0.9969

Accuracy: 0.9734

$$ Acc = \frac{TP}{Number of Testing Images}$$


Mean Recall: 0.8705

$$ {mRec} = {\rm mean}({Rec_i})$$


Mean Precision: 0.9163

$$ {mPre} = {\rm mean}({Pre_i})$$


Cohen Kappa Score: 0.9585

Confusion Matrix (%):

Predicted
Articulated Truck Bicycle Bus Car Motorcycle Non-motorized Vehicle Pedestrian Pickup Truck Single Unit Truck Work Van Background
True Articulated Truck 91.42 0.00 0.23 0.62 0.04 0.35 0.00 0.58 6.46 0.19 0.12
Bicycle 0.00 83.54 0.18 0.70 1.75 0.70 11.91 0.18 0.00 0.00 1.05
Bus 0.62 0.00 95.23 1.63 0.00 0.08 0.00 0.31 0.97 0.85 0.31
Car 0.01 0.00 0.01 98.57 0.00 0.01 0.00 1.21 0.01 0.13 0.05
Motorcycle 0.00 7.07 0.00 5.86 85.25 0.00 1.62 0.20 0.00 0.00 0.00
Non-motorized Vehicle 10.50 0.23 0.46 8.22 1.60 55.71 0.23 1.60 14.84 2.74 3.88
Pedestrian 0.00 1.60 0.00 0.32 0.19 0.06 95.85 0.00 0.00 0.00 1.98
Pickup Truck 0.03 0.00 0.02 6.21 0.04 0.06 0.00 93.23 0.26 0.16 0.01
Single Unit Truck 8.91 0.00 0.23 2.42 0.00 0.55 0.00 5.23 79.84 2.27 0.55
Work Van 0.17 0.00 0.25 16.35 0.00 0.25 0.04 2.27 1.49 79.19 0.00
Background 0.04 0.01 0.01 0.18 0.00 0.01 0.03 0.03 0.00 0.01 99.69