ShapeAug: Occlusion Augmentation for Event Camera Data (2401.02274v1)
Abstract: Recently, Dynamic Vision Sensors (DVSs) sparked a lot of interest due to their inherent advantages over conventional RGB cameras. These advantages include a low latency, a high dynamic range and a low energy consumption. Nevertheless, the processing of DVS data using Deep Learning (DL) methods remains a challenge, particularly since the availability of event training data is still limited. This leads to a need for event data augmentation techniques in order to improve accuracy as well as to avoid over-fitting on the training data. Another challenge especially in real world automotive applications is occlusion, meaning one object is hindering the view onto the object behind it. In this paper, we present a novel event data augmentation approach, which addresses this problem by introducing synthetic events for randomly moving objects in a scene. We test our method on multiple DVS classification datasets, resulting in an relative improvement of up to 6.5 % in top1-accuracy. Moreover, we apply our augmentation technique on the real world Gen1 Automotive Event Dataset for object detection, where we especially improve the detection of pedestrians by up to 5 %.
- A low power, fully event-based gesture recognition system. In Conference on Computer Vision and Pattern Recognition (CVPR).
- Object detection with spiking neural networks on automotive event data. In International Joint Conference on Neural Networks (IJCNN).
- A large scale event-based detection dataset for automotive. arXiv preprint arXiv:2001.08499.
- Imagenet: A large-scale hierarchical image database. In Conference on Computer Vision and Pattern Recognition (CVPR).
- Improved regularization of convolutional neural networks with cutout. arXiv preprint arXiv:1708.04552.
- Incorporating learnable membrane time constant to enhance learning of spiking neural networks. In International Conference on Computer Vision (ICCV).
- Learning generative visual models from few training examples: An incremental bayesian approach tested on 101 object categories. In Conference on Computer Vision and Pattern Recognition Workshop (CVPRW).
- Occlusions for effective data augmentation in image classification. In International Conference on Computer Vision Workshop (ICCVW).
- Recurrent vision transformers for object detection with event cameras. In Conference on Computer Vision and Pattern Recognition (CVPR).
- Spiking Neuron Models: Single Neurons, Populations, Plasticity. Cambridge University Press.
- Eventdrop: data augmentation for event-based learning. In International Joint Conferences on Artificial Intelligence (IIJCAI).
- Identity mappings in deep residual networks. In European Conference on Computer Vision (ECCV).
- Augment your batch: better training with larger batches. In arXiv.
- Densely connected convolutional networks. In Conference on Computer Vision and Pattern Recognition (CVPR).
- Intra-clip aggregation for video person re-identification. In International Conference on Image Processing (ICIP).
- Krizhevsky, A. (2012). Learning multiple layers of features from tiny images. University of Toronto.
- CIFAR10-DVS: An event-stream dataset for object classification. Frontiers in Neuroscience.
- Asynchronous spatio-temporal memory network for continuous event-based object detection. Transactions on Image Processing.
- Neuromorphic data augmentation for training spiking neural networks. In European Conference on Computer Vision (ECCV).
- Focal loss for dense object detection. Transactions on Pattern Analysis and Machine Intelligence (PAMI).
- SSD: Single shot MultiBox detector. In European Conference on Computer Vision (ECCV).
- Decoupled weight decay regularization. In International Conference on Learning Representations (ICLR).
- Converting static image datasets to spiking neuromorphic datasets using saccades. Frontiers in Neuroscience.
- Learning to detect objects with a 1 megapixel event camera. In Neural Information Processing Systems (NeurIPS).
- Eventmix: An efficient data augmentation strategy for event-based learning. Information Sciences.
- Hide-and-seek: Forcing a network to be meticulous for weakly-supervised object and action localization. In International Conference on Computer Vision (ICCV).
- Hats: Histograms of averaged time surfaces for robust event-based object classification. In Conference on Computer Vision and Pattern Recognition (CVPR).
- Dropout: A simple way to prevent neural networks from overfitting. Journal of Machine Learning Research (JMLR).
- CutMix: Regularization strategy to train strong classifiers with localizable features. In International Conference on Computer Vision (ICCV).
- Random erasing data augmentation. In AAAI conference on artificial intelligence.