Watch Serhii’s interview and learn more about facial recognition and CompreFace.
Types of Facial Recognition Solutions
If you are a business owner and you are considering integrating facial recognition into one of your systems, keep in mind that you’ll need to choose from different types of facial recognition solutions. There are four main categories of solutions currently on the market:
- Software as a service
Software as a service facial recognition solutions seem to be the easiest option, as they handle everything, and you don’t need to manage any servers. All you need to do is to integrate the solution into your current system, which is quite easy to do. These solutions usually have good documentation and provide clear and easy instructions. But the main disadvantage is that it is very expensive, especially if you have a high volume of requests. You may also have concerns about data security, since you’ll have to send sensitive information to a third party with minimal oversight.
- Paid self-hosted
Paid self-hosted solutions are the exact opposite; you deploy the solutions on your own servers (on the cloud or on-premises), which means that you always control your data. This option is also much cheaper than software as a service solutions. But since you’ll be managing your own servers, you’ll need to hire specialists.
- Free self-hosted/open source
There are also free self-hosted and open-source solutions, like Exadel CompreFace. The main advantage of this option is that you always control your data because you deploy it in your servers. You control everything; you’ll see the code and know what happens to incoming photos. And, since it’s open source, you can alter the code to fit your business. The only disadvantage is that you need to manage your own servers.
- A custom solution from scratch
Of course, you can always create your own solution. The advantages will sound familiar. You control your data and the program’s structure, so it will perfectly fit your business. But there are big disadvantages as well; you need to have very experienced developers, including experts in things like DevOps, machine learning, computer vision, architecture, databases, and more.
To summarize, here are the pros and cons of each type of solution:
Read the article on Toolbox by Serhii Pospielov to learn more about types of facial recognition solutions and their advantages and disadvantages.
Assessing the Cost of Facial Recognition Technology
Let’s look at an example of this in the real world by comparing Amazon’s Rekognition service and CompreFace. Rekognition is pretty straightforward; you pay $1 for every 1,000 requests. And if you don’t have many requests per year, you won’t pay much. But if you have a large application, you’ll end up paying a lot more, especially if you are using video recognition (like through a video stream).
Understanding CompreFace’s costs are a little more complicated, since the solution itself is free of charge, but you have to pay for the server. For this comparison, we estimated the cost of a rented server as $840 per year. Since this is a flat cost, your savings will increase as your request volume grows.
*This comparison is based on an estimation for the volume of requests and server rental fees
Read the article on Toolbox by Serhii Pospelov to explore more about the costs of facial recognition solutions and see the key business use cases of facial recognition technology
Facial Recognition Software Development with CompreFace
If you want to develop your facial recognition software, you should consider using CompreFace. CompreFace is free and open source, which means you can focus your budget entirely on having our team integrate and customize your product. The process looks like this:
Imagine that you need to develop an employee attendance system. The first thing you need is business logic; then you need a facial recognition service (for our purposes, CompreFace), which you’ll have to integrate into your system. Since CompreFace is free, you only pay for the development of your system and integration with the solutions.
Integrate and customize your facial recognition product cost-effectively
Exadel can help you use CompreFace to develop your own customized facial recognition product that will include:
- A client application — the customer-facing application that runs your whole face recognition product (can be web-based or mobile)
- An admin application — the side of the application that you’ll use to run your face recognition product
- Authorization — a system of roles that restricts some users from sensitive information
- Video/image processing server — this server will process incoming images (for example, a video processing server will split the video into frames/images that the solution can recognize)
- Other custom components — this could be almost anything depending on your business logic