← 学习库 Computer Vision: Algorithms and Applications (2nd) 本册目录

Chapter 5: Deep learning

• Small-scale datasets suitable for training a simple CNN as a useful teaching tool:¹ MNIST (LeCun, Cortes, and Burges 1998), CIFAR-100 (Krizhevsky 2009), and Fashion MNIST (Xiao, Rasul, and Vollgraf 2017).

  • PyTorch TorchVision provides a great way to easily download some of the popular computer vision datasets, https://pytorch.org/vision/stable/datasets.html. TensorFlow also provides similar support with TensorFlow Datasets, https://www.tensorflow.org/datasets.
原书第 771 页

C.1 Datasets and benchmarks

Widely used recognition, detection, and segmentation datasets and benchmarks, as listed in Tables 6.1–6.4; separate datasets for other tasks such as image enhancement, motion estimation, and stereo, are discussed in later sections.

← Chapter 4: Model fitting and optimizationChapter 6: Recognition →