Information
Contact Name: Carl Lemaire
Email: carl.lemaire@usherbrooke.ca
Organization: Université de Sherbrooke
Method's Name: Xception
Reference: Chollet, François. Xception: Deep Learning with Depthwise Separable Convolutions. arXiv preprint arXiv:1610.02357, 2016.
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
0.9278 0.8670 0.9752 0.9822 0.9281 0.7745 0.9409 0.9418 0.7625 0.8952 0.9978

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
0.8891 0.8564 0.9752 0.9872 0.9131 0.7763 0.9367 0.9202 0.8125 0.9067 0.9972

Accuracy: 0.9761

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


Mean Recall: 0.9064

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


Mean Precision: 0.9085

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


Cohen Kappa Score: 0.9627

Confusion Matrix (%):

Predicted
Articulated Truck Bicycle Bus Car Motorcycle Non-motorized Vehicle Pedestrian Pickup Truck Single Unit Truck Work Van Background
True Articulated Truck 88.91 0.00 0.50 0.27 0.00 0.93 0.00 0.70 7.92 0.50 0.27
Bicycle 0.00 85.64 0.18 0.88 3.50 0.18 9.28 0.00 0.00 0.00 0.35
Bus 0.27 0.00 97.52 0.97 0.04 0.35 0.00 0.16 0.16 0.54 0.00
Car 0.01 0.00 0.01 98.72 0.00 0.01 0.00 0.99 0.00 0.21 0.04
Motorcycle 0.20 2.42 0.00 3.23 91.31 0.20 0.81 0.00 0.40 0.00 1.41
Non-motorized Vehicle 5.94 0.46 1.60 1.83 0.68 77.63 0.46 0.46 6.16 1.60 3.20
Pedestrian 0.06 3.64 0.00 0.64 0.19 0.19 93.67 0.00 0.06 0.00 1.53
Pickup Truck 0.02 0.00 0.08 7.06 0.00 0.07 0.00 92.02 0.39 0.31 0.06
Single Unit Truck 9.45 0.08 0.78 0.63 0.08 2.11 0.00 2.34 81.25 3.05 0.23
Work Van 0.25 0.00 0.41 6.11 0.00 0.25 0.00 0.95 1.32 90.67 0.04
Background 0.02 0.01 0.01 0.10 0.01 0.04 0.08 0.00 0.01 0.01 99.72