2021 Hotel-ID 酒店识别数据集

2021 Hotel-ID 是一个用于酒店识别的数据集。该数据集旨在帮助打击人口贩卖。数据集可帮助追踪被带到相同酒店的当前和未来受害者。酒店识别是一项具有挑战性的细粒度视觉分类任务。该数据集由通过 TraffickCam 移动应用程序包上传的酒店房间图像组成。这些图像的质量与调查图像相似，调查图像往往质量低下、摄像角度不常见和高度遮挡，因此在这些图像上训练的模型很有可能准确缩小酒店范围，找到正确的相同酒店。

The 2021 Hotel-ID is a hotel recognition dataset intended to help combat human trafficking. It aids in tracking current and future victims who have been taken to the same hotel. Hotel recognition is a challenging fine-grained visual classification task. This dataset comprises hotel room images uploaded through the TraffickCam mobile application. The quality of these images aligns with that of investigative images, which are typically low-resolution, captured from uncommon angles, and highly occluded. Consequently, models trained on this dataset have a high likelihood of accurately narrowing down potential hotels and identifying the exact matching establishment.

搜集汇总

背景与挑战

背景概述

2021 Hotel-ID 酒店识别数据集是一个用于打击人口贩卖的图像分类数据集，包含通过TraffickCam应用上传的酒店房间图像，这些图像质量低、角度特殊且遮挡严重，旨在训练模型准确识别相同酒店以追踪受害者。数据集大小为24.4 GB，专注于细粒度视觉分类任务。