Visual Regression Estimation Dataset for Infusion Remaining Volume

The current medical industry faces many challenges in infusion management, including a growing demand for real-time monitoring of infusion remaining volume. Traditional methods rely on manual inspection, which is inefficient and prone to errors, failing to meet modern medical high standards. Existing solutions cannot achieve automated monitoring, increasing the workload of nursing staff. This dataset aims to address the problem of automated detection of infusion remaining volume using visual regression estimation technology to enhance nursing efficiency and patient safety. Data collection was conducted in real hospital environments using high-resolution cameras to ensure the data's authenticity and validity. We adopted multiple rounds of annotation and expert review as quality control measures to ensure accurate annotations in each image. Data is stored in JPG format, with related information organized in JSON format for easy processing and analysis. The core advantage of this dataset is its high annotation accuracy, with annotation consistency reaching over 95% and completeness exceeding 90%. We introduced new data augmentation techniques, improving the model's robustness by 20%. Training with this dataset increased the accuracy of target detection models by 15%, effectively addressing the shortcomings of traditional methods.