UniMiB SHAR
收藏资源简介:
UniMiB SHAR数据集是由米兰比可卡大学信息、系统和通信系创建的,旨在通过智能手机加速计数据进行人体活动识别和跌倒检测。该数据集包含11,771个样本,涵盖了30名年龄在18至60岁之间的参与者执行的日常活动和跌倒动作。样本分为17个细粒度类别,分为两个粗粒度类别:一类包含9种日常生活活动(ADL),另一类包含8种类型的跌倒。数据集设计时考虑了选择样本的各种标准,如ADL类型、年龄、性别等。该数据集已被用于评估四种不同的分类器和两种不同的特征向量,旨在为研究人员提供一个用于客观评估ADL识别和跌倒检测技术的基准。
The UniMiB SHAR dataset was created by the Department of Information, Systems and Communication of the University of Milan-Bicocca, aiming to conduct human activity recognition and fall detection using smartphone accelerometer data. This dataset contains 11,771 samples, covering daily activities and fall movements performed by 30 participants aged between 18 and 60 years old. The samples are categorized into 17 fine-grained classes, which are further divided into two coarse-grained categories: one consisting of 9 types of activities of daily living (ADL), and the other encompassing 8 types of falls. Various criteria for sample selection were taken into account during the dataset's design, including ADL types, age, gender, and other relevant factors. The dataset has been utilized to evaluate four distinct classifiers and two different feature vectors, with the goal of providing researchers with a benchmark for the objective assessment of ADL recognition and fall detection technologies.




