基于移动端的非侵入式贫血检测系统研究与设计

Research and Design of Non-Invasive Anemia Detection System Based on Mobile Terminals

摘要:

贫血是一种全球性健康问题,对身体有不良的影响。目前主流的贫血检测是通过采集血液样本来实现对贫血的检测,这种穿刺皮肤的方式可能会给个体带来一定的不适；因此提出了改进的ResNet34贫血检测算法,通过对残差块的改进和引入残差块组的整体跳跃连接有效地提高了分类准确率,并开发了一款非侵入式贫血检测的微信小程序,操作简单,使用便捷,可以实现随时随地的贫血检测,适应于用户的日常生活。在真实环境下的实验结果表明,该系统可以实现对测试样本的全部正确分类,其可用性和准确性基本满足用户需求,也为非侵入式贫血检测提供了有益的参考。

Abstract:

Anemia is a global health issue that adversely affects the body. Currently, the mainstream method for anemia detection involves collecting blood samples, which may cause discomfort due to skin puncture. Therefore, this article proposes an improved ResNet34 anemia detection algorithm. By enhancing residual blocks and introducing overall skip connections within residual block groups, the classification accuracy is effectively improved. Additionally, a non-invasive anemia detection WeChat mini-program is developed, featuring simple operation and convenient use, enabling anemia detection anytime and anywhere. It is suitable for users" daily lives. Experimental results in real-world settings demonstrate that the system achieves correct classification for all test samples. Its usability and accuracy largely meet user requirements, offering valuable insights for non-invasive anemia detection.