Medical video data on cloud infrastructures
Cloud computing is based on the paradigm of scalability and flexibility, allowing users to access resources remotely, without having to invest in their own IT infrastructure. Cloud services are based on the provider’s external infrastructure, i.e. Google Cloud Platform (GCP), Microsoft Azure and Amazon Web Services (AWS), etc.
Storage and processing data in the cloud have a huge number of benefits. I’ve already introduced a few of them. However, the main point of this post is to focus on a specific area: medical video data on cloud infrastructure. According to this, I would like to ask you a question: Have you ever wondered how endoscopic data is processed in the cloud and what it entails, also in terms of costs?
Let’s start with the process of uploading images from video to the cloud. Now, it can be seen as a sequence of operations. First, the video is divided into sections with a length adapted to the needs of a given project. Then, each section is divided into frames, which are compressed into image format. Saving images individually can be advantageous when there’s a frequent need to access or edit single frames, as often required in machine learning, without processing the entire video. However, this approach generally requires more storage space, making cloud solutions a desirable option.
Example of using cloud computing power for ML algorithm
Training an ML algorithm requires large amounts of data and computing power. As well as a properly prepared network infrastructure. The advantage of this solution is the availability and flexibility in the choice of tools and applications. Therefore, a properly prepared cloud computing allows you to create ML algorithms faster and more effectively.
Medical films capture dynamic processes such as labelling specific events, actions, or objects in video frames and tracking their movements or progress. It depends on what kind of results you care about or want to achieve.
Let’s take a closer look at this. We can illustrate it using the example of how a machine learning algorithm can be used to detect colorectal tumors from real-time endoscopic film data during a colonoscopy.
- Imaging: During endoscopy, an image of the inside of the colon is captured using an endoscopic camera.
- Upload to the cloud: The image is uploaded in real-time to the cloud computing.
- ML Analysis: A machine learning algorithm analyses the image and identifies areas suspected of colorectal cancer.
- Communicating information: The results of the ML analysis are sent back to the user, who can use them to investigate suspicious areas more closely.
Whereas the above example utilizes real-time data, AI’s ability to detect polyps or tumors from endoscopic camera footage offers a wider range of applications. It can be used not only for initial detection but also to monitor disease progression by analyzing past patient examinations (longitudinal study).
Brief summary: What kind of benefits brings storage medical video data on cloud?
There are many benefits to using the cloud to process data. You don’t have to invest in your server rooms, hardware, or software. Unlike traditional IT infrastructure, where you have to invest in servers and other devices that work even when they are not in use, the cloud allows you to optimize costs.
Speaking of costs, it is worth noting that we have the option of paying only for the cloud resources that we actually use at a given time. The so-called “pay-as-you-go” payment model allows you to pay only for the resources you actually use and only when they are running.
As you can see, cloud computing data offers many benefits. Besides efficiency and minimalized cost, it’s worth mentioning about security. Cloud-based service providers implement strict security measures such as encryption, access controls, and regular data backups, alleviating common concerns. In one of our latest posts, you can find out more about, how cloud computing influences the security of medical software.
To sum up, by using machine learning algorithms and cloud computing technologies, we can automate labour-intensive tasks, shorten the time of their analysis, and significantly reduce costs.