Determination of the location of the needle entry point based on an improved pruning algorithm

Area of Science:

• Medical Robotics
• Artificial Intelligence in Healthcare
• Image Processing

Background:

• The emergence of new coronaviruses and their variants has strained global medical resources.
• Reducing direct contact between healthcare professionals and patients is crucial for infection control.
• Automating procedures like intravenous injections can enhance safety and efficiency.

Purpose of the Study:

• To develop an intelligent robot for handback intravenous injections.
• To reduce the risk of infection transmission between medical staff and patients.
• To create core technologies for vein detection, segmentation, and optimal needle insertion point decision.

Main Methods:

• An improved U-Net mechanism (AT-U-Net) was developed for dorsal hand vein detection and segmentation.
• A self-built dorsal hand vein database was used for algorithm validation.
• An improved pruning algorithm (PT-Pruning) was proposed for determining the needle entry point, considering vascular characteristics.

Main Results:

• The AT-U-Net algorithm achieved an F1-score of 93.91% for vein detection and segmentation.
• The PT-Pruning algorithm achieved 96.73% accuracy in detecting effective injection areas.
• The system demonstrated 96.50% accuracy in identifying the optimal needle entry point.

Conclusions:

• The developed intelligent injection robot and its core algorithms provide a foundation for automated mechanical injection.
• The AT-U-Net and PT-Pruning algorithms show high accuracy and effectiveness in vein and injection point identification.
• This technology has the potential to significantly improve safety and efficiency in intravenous procedures.