目录
六、Convolution(卷积)与Cross-correlation(互关性)的区别
一、现实生活的CV的具体应用
1. Image classification(图像分类)
Input: Image
Output: class label
Datasets: MNIST、CIFAR10, CIFAR100、 ImageNet.
2. Object detection (目标检测)
Input: image
Output: object bounding boxes and class labels
Datasets: Pascal VOC 、MS-COCO.
3. Semantic segmentation (语义分割)
Input: image
Output: class label for each pixel
Datasets: