top of page

Rapid Vein Finder: VanTech serves as industry partner for ENG capstone team

1 Apr 2023

Making Intravenous Access Quicker, More Accurate and Affordable

The Need for Better Vein Detection Tools

Intravenous access is a routine, yet vital procedure in the healthcare sector. Whether its for resuscitation measures, hydration, administering medication, or preparing for medical imaging, a precise and efficient process is crucial. However, multiple attempts due to missed veins can lead to increased discomfort and stress for patients and healthcare professionals alike.

Vein imaging devices are gradually becoming the answer. But can they be made more accurate, cost-effective, and adaptable to varying conditions?

The answer is yes, and the solution is VanTech's M/R VEIN.

Introducing a Three-Phased Vein Detector

VanTech partnered with UBC Okanagan Engineering students (Caleb Kavaloff, Chase Anderson, Einar Schiele, Jesse Alele, Sathvik Radhakrishnan) on an initiative aiming to refine and rethink the vein detection process. The result was a machine-learning-integrated vein finder that promises accuracy on par with current market leaders, but at a fraction of the cost.

Phase 1: Template Matching

The initial phase revolved around template matching - a process where predefined vein templates are matched over a live image to detect veins. While promising, its dependency on fixed templates made it less adaptable to diverse skin tones and lighting conditions. The search was on for a more dynamic solution.

Phase 2: Real-time Video Processing

The answer lay in real-time video processing and contour detection. Here, an array of filters, transforms, and thresholds processed live footage to accentuate veins. The result was a system as efficient as current solutions, detecting veins effectively within a 5 to 10-inch distance from the patient.

Phase 3: Machine Learning - The Game-Changer

Finally, K-Means Clustering, an unsupervised machine learning technique, was introduced. Instead of relying on predefined templates, the system learned to detect veins by recognizing underlying patterns. Though the prototype needs further optimization, its potential is undeniable.

Why This is a Breakthrough

The final product is not just another vein detector. It is a multi-modal tool allowing users to pick from three distinct modes:

  1. Template Matching – Useful for straightforward cases, demonstrating the foundational technology.

  2. Video Processing – Offers quick and efficient vein detection, ideal for most standard conditions.

  3. Machine Learning – Although intensive, this mode demonstrates the future of vein detection, harnessing the power of AI.

At approximately $400 CAD for a prototype, this vein detector is significantly more affordable than its $5000+ USD counterparts, promising a revolution in patient care.

The Future is Bright

As the world of healthcare continues to evolve, so do the tools that aid its professionals. The future points towards AI-enhanced tools that not only make procedures more comfortable and efficient for patients but also more accessible for institutions. This vein detector is just the beginning!

bottom of page