加密视频、细粒度网页识别数据集

本数据集基于自动化采集与结构化处理方法，构建了覆盖多平台视频流量与细粒度网页识别数据。数据集包含两个核心部分：（1）多平台QUIC视频播放流量数据，包含YouTube、Facebook、Instagram三大平台的QUIC协议视频流量特征处理结果，同时记录了视频ID、时长、分辨率等相关元信息的；（2）细粒度网页识别数据，包含针对维基百科100个URL，100次重复访问得到的网页访问流量特征处理结果。所有数据采集均严格遵循《数据资源加工指导规范》，采用自动化脚本模拟真实用户交互行为，并通过协议解析验证确保数据时空精度。本数据集为QUIC场景下的视频识别、网页指纹识别算法开发、网络流量分类模型训练等研究提供基准数据支撑，具有重要的科研与工程应用价值。This dataset is constructed via automated collection and structured processing approaches, covering multi-platform video traffic and fine-grained web recognition data. The dataset consists of two core components: (1) Multi-platform QUIC video playback traffic data, which includes the processed feature results of QUIC protocol video traffic from three major platforms: YouTube, Facebook, and Instagram, while recording relevant metadata such as video ID, duration, and resolution; (2) Fine-grained web recognition data, which includes processed web access traffic feature results obtained from 100 repeated visits to 100 Wikipedia URLs. All data collection strictly complies with the Guidance Specifications for Data Resource Processing. Automated scripts are used to simulate real user interaction behaviors, and protocol parsing verification is adopted to ensure the spatiotemporal accuracy of the collected data. This dataset provides benchmark data support for research such as video recognition in QUIC scenarios, web fingerprint recognition algorithm development, and network traffic classification model training, and holds significant scientific research and engineering application value.

搜集汇总

背景与挑战

背景概述

该数据集包含多平台QUIC视频流量特征处理结果和细粒度网页访问流量特征处理结果，通过自动化采集与结构化处理构建。它为QUIC场景下的视频识别、网页指纹识别算法开发及网络流量分类模型训练提供基准数据支撑。