N-ImageNet: Towards Robust, Fine-Grained Object Recognition with Event Cameras (Mini Train/Validation Splits)

This repository contains N-ImageNet along with its variants. N-ImageNet is a large-scale dataset for event-based object recognition that also allows robustness evaluation in various external conditions. The train split is separated into 10 parts. To extract the event data in each .zip file, run the following command:

unzip train_Part_i.zip
for f in $( ls train_Part_i ); do tar -xvf train_Part_i/$f -C train_Part_i/ ;done
rm -rf train_Part_i/*.tar.gz

After this, make a separate folder extracted_train and extracted_val as specified in https://github.com/82magnolia/n_imagenet#dataset-setup and start using mini N-ImageNet!