5 min readingIn Part I and Part II of this series, we discussed how to navigate the build vs. buy dilemma as a digital health start-up and applied this decision-making framework to computer vision technology. So let’s say you’ve decided to license computer vision technology rather than building it in-house. How do you decide on the right vendor with the technology that is right for you?
Let’s find out!
When evaluating the computer vision that is right for you, there are several key criteria to consider:
- Customization Ability
- Integration Options
- Scalability
- Regulatory Compliance
and, perhaps, most importantly…
- Technology Performance and Sophistication
Customization Ability
Yes, you’ve decided that it makes the most sense to license existing technology rather than building it yourself, but it’s still important to make sure that the technology can be customized to fit your use case. Computer vision solutions can appear very attractive from simple demos, but is there advanced functionality behind the polished demo? For example, it’s important to understand branding and white-labeling options and the availability of various UX configurations (e.g., display settings, layout options).
Integration Options
Of course, the power of the computer vision technology doesn’t matter if it can’t integrate effectively into your organization’s workflow and UX. It’s important to have a detailed discussion with the vendor about the technical details before making a decision, and make sure to pull your own tech team into this discussion. Does the vendor offer an API framework or just a standalone solution? Does the organization offer a modular approach to integration, or is it a one-size-fits-all solution? Can the solution integrate on mobile devices and desktops? Are the device requirements restrictive? These are all critical questions to know the answer to before making a decision
Scalability
Hopefully, your business will continue to grow, and you must make sure that the computer vision solution they select is equipped to scale with your digital health start-up. This involves a variety of product-focused considerations, including the expansion of use cases and growing data volumes. If your organization expands use cases to support more conditions or different types of users, will the computer vision be able to adapt? Does the computer vision provider capture the granularity of data you’ll need as your user population grows? Is there any hardware needed to support the product that will limit scalability at higher volumes? In addition to these product considerations, it’s necessary to also consider scalability from a business perspective. For example, you should understand whether the vendor’s pricing model will align with your business needs and enable financial benefits as you grow. Ultimately, the computer vision technology you integrate should make sense with your vision, not just your current situation.
Regulatory Compliance and Privacy
In healthcare, regulatory compliance is always a key consideration, regardless of the country or patient population. While you may devote significant resources and energy to ensuring that your product and platform is compliant with all the relevant regulations, not all computer vision vendors are as diligent and compliant. Ensure that the vendor you select meets the requirements (e.g., HIPAA, FDA, GDPR) for each of your markets. Additionally, privacy is an important concern in the computer vision space because of the camera usage. Make sure that patient or user data will be kept secure and that the computer vision technology does not infringe upon your users’ privacy.
Technology Performance and Sophistication
The performance and sophistication of the computer vision technology cannot be separated from any of the above criteria. A solution that misses the mark here will also miss the mark with regards to customization, integration, scalability, and compliance and privacy. However, because computer vision technology is relatively new for many in the healthcare industry, it can be difficult to understand what is important and what might not be as relevant. It can be helpful to develop a list of the performance criteria you are looking for and then evaluate options using that list.
For instance, if you’re looking to use computer vision for exercising and human motion evaluation, these are the key dimensions of performance and sophistication:
- Precise and complete body tracking
- Real-time feedback and corrective guidance
- Diverse exercise library with thousands of exercises in various positions
- Accurate adherence tracking that includes actual exercise completion details
- Robust API framework for seamless integration into other digital platforms
- Granular performance tracking at the exercise level
- Wide support of devices
If you need more information on how to evaluate whether a solution meets these criteria, reach out to info@kemtai.com, and we can share a simple solution that clearly outlines the core components within each criteria. Of course, we may be a bit biased, but we’d be happy to help you better understand what computer vision is right for you.
