1. CIFAR-10 & CIFAR-100
CIFAR-10包含10个类别,50,000个训练图像,彩色图像大小:32x32,10,000个测试图像。
(类别:airplane,automobile, bird, cat, deer, dog, frog, horse, ship, truck)
(作者:Alex Krizhevsky, Vinod Nair, and Geoffrey Hinton)
(数据格式:Python版本、Matlab版本、二进制版本<for C程序>)
CIFAR-100与CIFAR-10类似,包含100个类,每类有600张图片,其中500张用于训练,100张用于测试;这100个类分组成20个超类。每个图像有一个"find" label和一个"coarse"label。
2. 图像分类结果及对应的论文
图像分类结果及应的论文,包含数据集:MNIST、CIFAR-10、CIFAR-100、STL-10、SVHN、ILSVRC2012 task 1
ILSVRC: ImageNet Large Scale Visual Recognition Challenge
3. ImageNet
ImageNet相关信息如下:
1)Total number of non-empty synsets: 21841
2)Total number of images:
14,197,122
3)Number of images with
bounding box annotations: 1,034,908
4)Number of synsets with SIFT features: 1000
5)Number of images with SIFT features: 1.2
million
4. COCO
COCO(Common Objects in Context)是一个新的图像识别、分割、和字幕数据集,它有如下特点:
1)Object segmentation
2)Recognition in Context
3)Multiple objects per image
4)More than 300,000 images
5)More than 2 Million instances
6)80 object categories
7)5
captions per image
8)Keypoints on
100,000 people
COCO 2016 Detection Challenge(2016.6.1-2016.9.9)和COCO 2016 Keypoint Challenge(2016.6.1-2016.9.9)已经由Microsoft发起 由ECCV 2016(ECCV:European Conference On Computer Vision )。
4. 3D数据
3)Human3.6M (3D Human Pose Dataset)
- 《Iterated Second-Order Label Sensitive Pooling for 3D Human Pose Estimation》
5. 人脸Dataset
1)LFW (Labeled Faces in the Wild)
6. Stereo Datasets
3)KITTI Vision Benchmark Suite