CURE-OR (Challenging Unreal and Real Environments for Object Recognition)

CURE-OR是一个大规模，受控和多平台的对象识别数据集，被表示为挑战对象识别的虚幻和真实环境。在该数据集中，存在具有不同大小，颜色和纹理的100对象的1,000,000图像，这些图像以五个不同的方向定位，并使用五个设备 (包括网络摄像头，DSLR和三个智能手机相机) 在真实世界 (真实) 和工作室 (虚幻) 环境中捕获。受控的挑战性条件包括曝光不足、曝光过度、模糊、对比度、脏镜头、图像噪声、调整大小和颜色信息丢失。

CURE-OR is a large-scale, controlled, and multi-platform object recognition dataset developed to challenge object recognition tasks in both virtual and real-world environments. This dataset contains 1,000,000 images of 100 distinct objects with varying sizes, colors, and textures. These objects are positioned in five different orientations, and the images are captured across real-world (real) and studio (virtual) environments using five devices, including webcams, DSLRs, and three smartphone cameras. The controlled challenging conditions include underexposure, overexposure, blur, contrast variations, lens dirt, image noise, resizing, and color information loss.

搜集汇总

背景与挑战

背景概述

CURE-OR是一个大规模、受控且多平台的对象识别数据集，包含100个不同对象的100万张图像。这些图像在真实和虚幻环境中使用五种设备捕获，并涵盖了多种挑战性条件，如曝光异常、模糊和噪声等。