SIP (Salient Person) dataset

SIP数据集是由南开大学计算机学院创建的，包含约1K高分辨率图像，涵盖了多样化的真实世界场景，如不同的视角、姿势、遮挡、光照和背景。该数据集特别关注人类在视觉场景中的显著性检测，旨在为移动设备上的RGB-D模型应用提供支持。数据集中的图像经过精心设计，以覆盖多种挑战性情况，并提供了像素级的真实标注。此外，SIP数据集还提供了RGB和灰度图像，这些图像由双目相机捕获，有助于多种研究方向，如立体匹配、深度估计和以人为中心的检测等。数据集的应用领域包括提高智能手机摄像头的性能，以及在监控探头和工业机器人中的应用，旨在解决利用深度信息进行显著对象检测的问题。

The SIP Dataset was developed by the College of Computer Science at Nankai University. It contains approximately 1,000 high-resolution images covering diverse real-world scenarios, including varying viewpoints, poses, occlusions, lighting conditions and backgrounds. This dataset specifically focuses on human-centric visual saliency detection, aiming to support the deployment of RGB-D models on mobile devices. The images in the dataset are meticulously designed to cover a wide range of challenging situations, and are accompanied by pixel-level ground truth annotations. Additionally, the SIP Dataset provides both RGB and grayscale images captured by a binocular camera, which can facilitate multiple research directions such as stereo matching, depth estimation and human-centered detection. The application fields of this dataset include enhancing the performance of smartphone cameras, as well as applications in surveillance cameras and industrial robots, with the objective of addressing the problem of salient object detection using depth information.