CIFAR10-DVS: An Event-Stream Dataset for Object Classification

By Hongmin Li, Hanchao Liu, Xiangyang Ji, Guoqi Li, Luping P. Shi. Frontiers in Neuroscience, 2017

710 citations recorded by the source corpus.

Neuromorphic vision research requires high-quality and appropriately challenging event-stream datasets to support continuous improvement of algorithms and methods. However, creating event-stream datasets is a time-consuming task, which needs to be recorded using the neuromorphic cameras. Currently, there are limited event-stream datasets available. In this work, by utilizing the popular computer vision dataset CIFAR-10, we converted 10,000 frame-based images into 10,000 event streams using a dynamic vision sensor (DVS), providing an event-stream dataset of intermediate difficulty in 10 different classes, named as “CIFAR10-DVS.” The conversion of event-stream dataset was implemented by a repeated closed-loop smooth (RCLS) movement of frame-based images. Unlike the conversion of frame-based images by moving the camera, the image movement is more realistic in respect of its practical applications. The repeated closed-loop image movement generates rich local intensity changes in continuous time which are quantized by each pixel of the DVS camera to generate events. Furthermore, a performance benchmark in event-driven object classification is provided based on state-of-the-art classification algorithms. This work provides a large event-stream dataset and an initial benchmark for comparison, which may boost algorithm developments in even-driven pattern recognition and object classification.

Subfields: Cellular and Molecular Neuroscience, Electrical and Electronic Engineering

View Paper: https://www.semanticscholar.org/paper/e72b7962133921fa3e84299cd6a4a2aeb60bab19

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Cellular and Molecular Neuroscience

1. Yonghong Tian
2. Lei Deng
3. Guoqi Li
4. Zhaofei Yu
5. Ryad Benosman
6. Xavier Lagorce
7. Wei Fang
8. Yanqing Chen

In Electrical and Electronic Engineering

1. Yonghong Tian
2. Guoqi Li
3. Timothée Masquelier
4. Zhaofei Yu
5. Wei Fang
6. Yanqing Chen
7. Liuliang Yuan
8. Man Yao

Citations in the corpus

• Going Deeper With Directly-Trained Larger Spiking Neural Networks (2020, methodology): “We test our deep residual SNNs on static non-spiking datasets (CIFAR-10, ImageNet) and neuromorphic datasets (DVS-gesture, DVS-CIFAR10).”
• Spike-driven Transformer (2023, methodology): “The latter contains CIFAR10-DVS [78] and DVS128 Gesture [79].”
• Deep Residual Learning in Spiking Neural Networks (2021, methodology): “We evaluate Spiking ResNet and SEW ResNet on both the static ImageNet dataset and the neuromorphic DVS Gesture dataset [1], CIFAR10-DVS dataset [32].”
• Anomaly Detection using One-Class Neural Networks (2018, background): “We observe that for MNIST and USPS datasets, which have less complicated spatial structures than CIFAR-10 images [25], the OC-SVM also produces best results.”
• Incorporating Learnable Membrane Time Constant to Enhance Learning of Spiking Neural Networks (2020, methodology): “The winner-take-all mechanism in the spatial domain and time-variant topology in the temporal domain achieved by max-pooling can increase SNNs’ fitting capability in temporal tasks, such as classifying the CIFAR10-DVS dataset.”
• Temporal Efficient Training of Spiking Neural Network via Gradient Re-weighting (2022, background): “Here, we adopt VGGSNN on the DVS-CIFAR10 dataset, set λ = 0 .”
• HATS: Histograms of Averaged Time Surfaces for Robust Event-Based Object Classification (2018, methodology): “Event-based Datasets An issue of previous work on event-based object classification is that the proposed solutions are tested either on very small datasets [64, 30], or on datasets generated by converting standard videos or images to an event-based representation [48, 23, 35].”