MedVidCL (Medical Video Classification)

MedVidCL 数据集包含 6、617 个视频的集合，这些视频被注释为“医学教学”、“医学非教学”和“非医学”类。使用两步法构建 MedVidCL 数据集。第一步，由健康信息学专家注释的视频用于训练机器学习模型，该模型将给定视频预测为上述三个类别之一。在第二步中，仅使用高置信度视频，健康信息学专家评估模型的预测视频类别并在需要时更新类别。

The MedVidCL dataset is a collection of 6,617 videos annotated into three categories: "Medical Teaching", "Medical Non-Teaching", and "Non-Medical". The MedVidCL dataset was constructed using a two-stage approach. In the first stage, videos annotated by health informatics experts were utilized to train a machine learning model that can classify a given video into one of the three aforementioned categories. In the second stage, only videos with high confidence scores were selected, and health informatics experts evaluated the category predictions generated by the model and updated the labels when necessary.

搜集汇总

背景与挑战

背景概述

MedVidCL是一个医疗视频分类数据集，包含6,617个视频，标注为'医学教学'、'医学非教学'和'非医学'三类。它采用两步法构建：先由专家标注视频训练分类模型，再对高置信度预测进行人工评估和修正。