ebdataset
收藏资源简介:
一个基于numpy记录数组的稀疏表示和PyTorch的密集表示构建的事件驱动数据集加载器,支持多种数据集,如Neuromorphic Mnist、NCaltech101等。
An event-driven dataset loader constructed based on sparse representations using numpy record arrays and dense representations using PyTorch, supporting various datasets such as Neuromorphic Mnist, NCaltech101, and more.
数据集概述
支持的数据集
-
Neuromorphic Mnist (N-MNIST)
- 来源:Orchard, G.; Cohen, G.; Jayawant, A.; and Thakor, N.
- 描述:“Converting Static Image Datasets to Spiking Neuromorphic Datasets Using Saccades”
- 出版:Frontiers in Neuroscience, vol.9, no.437, Oct. 2015
- 下载链接:N-MNIST
-
NCaTech101
- 来源:Orchard, G.; Cohen, G.; Jayawant, A.; and Thakor, N.
- 描述:“Converting Static Image Datasets to Spiking Neuromorphic Datasets Using Saccades”
- 出版:Frontiers in Neuroscience, vol.9, no.437, Oct. 2015
- 下载链接:NCaTech101
-
IBM DVS128 Gesture
- 来源:A. Amir, B. Taba, D. Berg, T. Melano, J. McKinstry, C. Di Nolfo, T. Nayak, A. Andreopoulos, G. Garreau, M. Mendoza, J. Kusnitz, M. Debole, S. Esser, T. Delbruck, M. Flickner, and D. Modha
- 描述:“A Low Power, Fully Event-Based Gesture Recognition System”
- 出版:2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR)
- 下载链接:IBM DVS128 Gesture
-
INI Roshambo17
- 来源:I.-A. Lungu, F. Corradi, and T. Delbruck
- 描述:“Live Demonstration: Convolutional Neural Network Driven by Dynamic Vision Sensor Playing RoShamBo”
- 出版:2017 IEEE Symposium on Circuits and Systems (ISCAS 2017)
- 下载链接:INI Roshambo17
-
INI UCF-50
- 来源:Hu, Y., Liu, H., Pfeiffer, M., and Delbruck, T.
- 描述:“DVS Benchmark Datasets for Object Tracking, Action Recognition and Object Recognition”
- 出版:Front. Neurosci. 10, 405. doi:10.3389/fnins.2016.00405
- 下载链接:INI UCF-50
-
NTidigits
- 来源:Anumula, Jithendar, et al.
- 描述:“Feature Representations for Neuromorphic Audio Spike Streams”
- 出版:Frontiers in Neuroscience, vol. 12, Feb. 2018, p. 23
- 下载链接:NTidigits
-
Prophesee N-Cars
- 来源:Amos Sironi, Manuele Brambilla, Nicolas Bourdis, Xavier Lagorce, Ryad Benosman
- 描述:“HATS: Histograms of Averaged Time Surfaces for Robust Event-based Object Classification”
- 出版:To appear in IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2018
- 下载链接:Prophesee N-Cars
数据集使用示例
- NMnist数据集
-
稀疏表示使用示例: python from ebdataset.vision import NMnist for spike_train, label in NMnist(path): spike_train.x, spike_train.y, spike_train.p, spike_train.ts break
-
密集表示使用示例: python from ebdataset.vision.transforms import ToDense dt = 1*ms loader = NMnist(path, is_train=True, transforms=ToDense(dt)) for spike_train, label in loader: spike_train.shape # => (34, 34, 2, duration of sample) break
-
可视化示例: bash python -m ebdataset.visualization.spike_train_to_vid NMnist path
-
安装指南
- 安装命令: bash pip install ebdataset




