VOC-360 和 Wider-360
收藏资源简介:
VOC-360和Wider-360是西蒙弗雷泽大学工程科学学院创建的两个新型鱼眼图像数据集,用于训练面部和物体检测模型。VOC-360包含39,575张用于物体检测、分割和分类的鱼眼图像,而Wider-360包含63,897张用于面部检测的鱼眼图像。这些数据集通过将VOC2012和Wider Face数据集中的常规图像转换为鱼眼图像而创建,适用于开发鱼眼图像的面部和物体检测器及分割模块。数据集的应用领域包括计算机视觉和模式识别,旨在解决鱼眼图像中的物体检测和面部识别问题。
VOC-360 and Wider-360 are two novel fisheye image datasets developed by the School of Engineering Science at Simon Fraser University, specifically for training facial and object detection models. VOC-360 comprises 39,575 fisheye images designed for object detection, segmentation and classification tasks, whereas Wider-360 contains 63,897 fisheye images tailored for face detection. These datasets are constructed by converting standard images from the VOC2012 and Wider Face datasets into fisheye imagery, and serve as valuable resources for developing facial and object detectors as well as segmentation modules for fisheye images. Their applicable domains cover computer vision and pattern recognition, and they are intended to address the core challenges of object detection and face recognition in fisheye image scenarios.
- 1Datasets for Face and Object Detection in Fisheye Images西蒙弗雷泽大学工程科学学院 · 2019年



