icarofua / vehicle-rear

Vehicle-Rear: A New Dataset to Explore Feature Fusion For Vehicle Identification Using Convolutional Neural Networks
Apache License 2.0
111 stars 19 forks source link

How to convert data from videos to data? #11

Closed woshiwoyali closed 4 years ago

woshiwoyali commented 4 years ago

Hello, how to convert the data source in videos to the data format in the data directory?

icarofua commented 4 years ago

Hi. I extracted from the xmls file the frame number and first plate position. After, I used a mosse tracker for each vehicle and extract until 10 images of each vehicle or until when the psr is low.