Keras Tutorial – Traffic Sign Recognition

chsasank.github.io

186 points·sytelus··9 comments

I like this example (especially as the dataset is simple, but new to me).

For something simpler, starting from very shallow networks (and expanding them step by step), I wrote "Starting deep learning hands-on: image classification on CIFAR-10" https://blog.deepsense.ai/deep-learning-hands-on-image-class....

This is a nice toy application. Yay .. something other than MNIST! Is there a similar toy application for RNNs/LSTMs? I don't find the standard text generating application to be compelling. I've been looking for a simple dataset for teaching/learning purposes ... something like accelerometer or acoustic data. Haven't found a great one yet for recurrent networks.

I enjoyed experimenting with RNNoise for real-time microphone noise suppression. I liked the write-up too [1].

[1]: https://people.xiph.org/~jm/demo/rnnoise/

EDIT: Just realized perhaps you are asking for a Keras-based example? If so, this one's not Keras or TF based.

Hey! Author here. May be you'd appreciate this: https://chsasank.github.io/spoken-language-understanding.htm...

This is neat. Thanks so much!

The article made me think of that it's still a long way to a adaptive classifier that could be shown one of two examples instead of thousands.

Can we get rid of the traffic sign image captcha now machines can do it?

You missed the point.. I think the traffic sign captchas are actually used to train machine learning models.

Yup, just wait until somebody implements a SSD multibox plugin for captchas...

Reply on news.ycombinator.com