Skip to content

SajjadPSavoji/German-Traffic-Sign-Recognition-Benchmark

Repository files navigation

GTSRB

project banner

The German Traffic Sign Benchmark is a multi-class, single-image classification challenge held at the International Joint Conference on Neural Networks (IJCNN) 2011. In this case study we examine the effect of common components in CNNs(including normalization layers) on the traceability and generalization capabilities

Resutls

base model

Model Accuray Loss
Base Model

activation functions

activation functions accuracy

optimizaton method

optimization method

drop out layers

Model with Drop out no Drop out
-----

data augmentation

dataaugmentation

batch normalization

Model accuracy loss
Batch Norm

Resources