SCO dataset : A dataset for semantic communication

The integration of communication and artificial intelligence has become a development trend, one of the applications is semantic communication, but the current research lacks the support of comprehensive datasets. To solve this problem, we built a new image and video dataset, named SCO dataset, for the researches on semantic communication and computing. First, we introduce the peculiarities of the dataset, which contains 5100 images and 138 video clips. Secondly, we we give the data generation and processing methods of the dataset, including images and videos. Then we describes the labeling method of the dataset for intelligent tasks, including classfication and object detection tasks, and the evaluation method of the dataset for subjective experience of human users. Next, we give and analysize the experimental results of the dataset. Finally, we give some practical applications of this dataset, including end-to-end semantic communication, joint source and channels semantic coding, and multi-user semantic communication resource management.