Child Pneumonia Dataset

Title: Child Pneumonia X-ray Dataset: A High-Quality Medical Imaging Resource for AI-Driven Diagnosis Authors: MD. Abu Raihan, Syed Muntasin Fayaz, Aziza Haque, Afia Sarkar. Affiliations: Department of Computer Science and Engineering, Khwaja Yunus Ali University, Sirajganj, Bangladesh . Description: A chest X-ray is the most popular and widely recognized way to diagnose pneumonia. We made use of a collection of medical chest X-ray pictures that were obtained from Sirajganj, Bangladesh's North Bangle Medical Hospital and College. These pictures show inflammation of the lungs brought on by bacteria, viruses, or chemicals that irritate the lungs. A total of 282 X-ray images were collected from the various mobile devices used to take the pictures. We personally examined and eliminated photos with distracting backgrounds, low quality, motion blur, or other artifacts to guarantee high-quality data. In order to preserve the integrity of the dataset, we also carried out a thorough quality check procedure, removing any photos with high brightness, poor contrast, or fuzzy areas. Subject Areas: Computer Sciences, Heath Science and Medical Science, Artificial Intelligence, Computer Vision and Pattern Recognition, Medical Imaging. Data Format: JPG images (processed and filtered) Data Collection: The devices used for capturing the images were the Xiaomi Redmi Note 7 and the Xiaomi Redmi Note 12. All raw images were captured under optimal lighting conditions and stored in JPG format. We portioned our dataset into 80:20 ratio for Training and Testing.标题：儿童肺炎X射线数据集：面向人工智能辅助诊断的高质量医学影像资源 作者：MD. Abu Raihan、Syed Muntasin Fayaz、Aziza Haque、Afia Sarkar 所属机构：孟加拉国锡拉杰甘杰市Khwaja Yunus Ali大学计算机科学与工程系 数据集描述：胸部X射线检查是目前临床诊断肺炎最常用且认可度最高的影像学方法。本研究采集的胸部X线影像均来自孟加拉国锡拉杰甘杰市北孟加拉医学院附属医院，涵盖了由细菌、病毒或肺部刺激物引发的肺部炎症病例。研究团队共收集了282张由不同移动设备拍摄的X线影像。为保障数据质量，我们人工筛查并剔除了背景杂乱、画质低劣、运动模糊或存在其他伪影的图像；同时开展了全面的质量校验流程，移除了亮度异常、对比度不佳或存在模糊区域的影像，以确保数据集的完整性与可靠性。 研究领域：计算机科学、健康科学与医学、人工智能、计算机视觉与模式识别、医学影像学 数据格式：JPG格式图像（已完成处理与筛选） 数据采集：本数据集的影像采集设备为小米Redmi Note 7与小米Redmi Note 12。所有原始图像均在最佳光照条件下拍摄，并以JPG格式存储。我们将数据集按照80:20的比例划分为训练集与测试集。