Interview with Bassam Shoukri and Willem de Fiouw from Business Data Challengers about the development of the BDC Image Classification Tool
Bassam Shoukri, one of the developers of BDC's new product, answers some questions about the development and its possibilities.
1. You presented a new tool this week. The BDC Image Classification Tool. How and when did the idea to develop this come about?
Together with Hetty Wessel, I took the initiative to explore with a number of colleagues the possibilities of using our knowledge and experience in the field of artificial intelligence, machine learning, data science, and processes to develop innovative solutions with added value for the customer.
In November 2019, we came up with the idea of using artificial intelligence for MRI image classification. The impact of this application has a major influence on the working methods within the domain of healthcare and medicine.
2. Who or what inspired you to do this?
In the BDC innovation club we think about possible new solutions and new developments. The goal is that the solution or idea benefits the community and the business. Our knowledge of the latest technology and the added value of this technology for the client organization form the starting points for this. From this approach we discovered how image classification is very valuable for the healthcare sector, among others.
3. Do you work with medical specialists?
We work together with Bytesnet Data Center in Groningen and also have a lot of contact with specialists in various hospitals to see how we are going to apply this using anonymized data from these institutions.
4. Do you have an estimate of how much time this could save medical specialists?
The use of artificial intelligence saves medical specialists a lot of time in classifying MRI images. This solution helps medical specialists to look at the images in a more targeted way by alerting the specialist to specific parts of the images that require extra attention. An additional advantage is the speed of classification. This method saves a lot of time because a large number of images can be classified in a very short time instead of having the images assessed one by one by a medical specialist.
5. Do you plan to develop more of these tools for healthcare? If so, can you tell us something about them?
In the BDC innovation club, we are always thinking about new technologies and advanced processes to improve our customer service.
In healthcare, the Image Classification tool can also be used to classify blood samples. BDC tested this with malaria-infected blood images. But the tool can also be used to detect abnormalities on x-rays or to detect Covid-19 in a blood image.
The product is flexibly developed. With minimal adjustments, the Image Classification tool can be used in other domains. This makes it possible to adapt the technology and use it for object recognition.
At the moment we have started a next innovation project, the use of text mining and NLP (Natural Language Processing). With this we help organizations to make analyses that are based on large amounts of text and documentation. Translate content text into a value in a dashboard. A complex whole that with the use of new technology delivers a lot of value for our customers.
Also this question to Willem de Fiouw:
Are there other areas where Business Data Challengers will offer innovative solutions?
The healthcare market is an important sector for us. We believe that the use of technology makes an important contribution to more effective and efficient and therefore affordable healthcare. By developing the Data Health scan, the Image Classification tool but also our solution for the capacity problem in healthcare, we are fulfilling our ambition to offer socially involved services.
In addition to the healthcare market, the security domain is an important segment for us where we have a lot of knowledge and experience. One of the innovative initiatives that we are taking up with the market concerns a solution to use AI and ML to extract required information from many thousands of texts and documents and to label them correctly based on the content of the text and then present this in clear dashboards.
Business Data Challengers
