Image Annotation
Human & Animal Blood Cell Image Annotation
The Problem
An award-winning healthcare AI company specializing in using the best in human and computational intelligence approached DataForce to build an external team with a background in biology to review and annotate thousands of microscopic images of blood smear samples. With 30,000 images across six species, our client was looking to deliver precise and individualized healthcare while also ridding the world of preventable diseases through rapid at-home testing for immune health. To accomplish this goal, artificial general intelligence was applied to mimic and enhance human medical reasoning to monitor individual users’ health for their at-home testing solution. Once the data was annotated, the machine learning model would be trained and equipped with the labeled data needed for the algorithm to operate efficiently and effectively.
• • • •The Solution• • • •
With a local presence in the UK, DataForce was able to begin the project within five days, sourcing students, graduates, and professionals with a degree in biology, hematology, and other related fields. First, the applicable participants reviewed the species by type, including cat, cow, dog, goat, horse, sheep, and human. From there, the participants began annotating the data, drawing bounding boxes around the relevant cells and identifying the correct white cell type. As the project concluded, the total number of images grew to about 40,000, exceeding the scope by nearly 10,000. The number of annotated cells surpassed 235,000, with an average of 6.014 cells per image, varying between animal and human. With this data, our client was able to refine their mission to deliver precise and individualized healthcare to their patients.
DataForce has a global community of over 1,000,000 members from around the globe and linguistic experts in over 250 languages. DataForce is its own platform but can also use client or third-party tools. This way, your data is always under control.