多源智能违禁品检测信息系统数据集

围绕传统安检系统中存在的数据源单一、信息孤立、图像匹配困难、小目标漏检、识别鲁棒性差、抽检效率低、多中心协同不足及高危物品识别难等关键问题，提出了一套完整的多源多模态安检数据智能处理算法框架，采集相关图片数据进行检测与验证，形成相关数据集。数据量676MB。

To address the key issues existing in traditional security inspection systems, including single data source, isolated information, difficulty in image matching, missed detection of small targets, poor recognition robustness, low sampling efficiency, insufficient multi-center collaboration, and difficulty in identifying high-risk items, this study proposes a comprehensive intelligent processing algorithm framework for multi-source and multi-modal security inspection data. Relevant image data are collected for detection and validation to form the corresponding dataset, with a total size of 676 MB.

搜集汇总

背景与挑战

背景概述

该数据集是针对传统安检系统中的多源多模态违禁品识别问题，提出了一套智能处理算法框架，并采集了相关图片数据进行验证。数据集包含676.56MB的16个文件，主要格式为zip和pdf。