Using Thermography and Medical Expert Annotations for Home Diagnostic AI Application
Kelvin Health
Industry
Medical
Type of service
Semantic Segmentation
Platform used
Results
Thermal and Medical images annotation for Kelvin Health
Kelvin Health, a prominent healthcare company focused on developing a groundbreaking home diagnostic application, sought to revolutionise the detection of vascular conditions, inflammatory diseases, and oncology-related anomalies using thermography.
In the fast-paced world of healthcare technology, accurate and reliable data annotation plays a pivotal role in advancing diagnostic applications. Humans in the Loop recently completed a remarkable project for Kelvin Health. This case study highlights our commitment to excellence, collaboration, and leveraging medically experienced professionals to achieve success.
The challenge
High-quality thermal medical image annotation and flexibility
With thermal imaging as the foundation, their AI-powered application developed by Kelvin Health required high-quality annotated data for accurate segmentation and precise identification of abnormalities. However, the challenge lay for them in finding a data annotation partner with the necessary medical expertise and flexibility to iterate alongside them during the project of thermal medical image annotation.
Tasked with segmenting over 2,000 thermography images for their AI-powered home diagnostic application, Humans in the Loop overcame unique challenges, maintained effective communication, and delivered exceptional results within a strict deadline.
We already have the first results and I must say they are amazing! Kudos to Mohamed and his team for the great medical insights and hard work put into the annotations. Thank you, all HITL-ers, for your commitment in this project.
Georgi Kostadinov,
Co-founder & CTO, Kelvin Health
The solution
Recognizing the unique demands of thermal medical image annotation for the project, Humans in the Loop assembled a highly skilled team consisting of certified medical professionals and experienced data annotators. To streamline the annotation process, Humans in the Loop utilised the powerful Supervise.ly platform, which provided a comprehensive suite of annotation tools and collaborative features. Leveraging our domain knowledge and the platform’s capabilities, we devised a meticulous workflow that ensured accurate segmentation of body thermal images, enabling the AI model to effectively identify anomalies.
The foremost challenge in this project was the requirement for medically experienced professionals to perform the annotations. Understanding the criticality of accurate medical annotations, we involved one of our certified medical doctors to oversee the project and guarantee the highest level of quality assurance. By combining the expertise of our medical professionals with the skills of our data annotators, we maintained the utmost precision in marking vascular conditions, inflammatory diseases, and oncology-related anomalies within the thermography images.
Additionally, our client emphasised the importance of collaboration and responsiveness throughout the project. Recognizing this need, we established a proactive and transparent communication channel, ensuring that our client was constantly informed about the progress and any issues that arose. Humans in the Loop prioritised implementing client feedback promptly, allowing for iterative improvements during the annotation process. This collaborative approach created a strong partnership that fostered trust and yielded outstanding results.
The result
Through a combination of medically experienced professionals, the utilisation of the Supervise.ly platform, and a collaborative approach with our client, HITL successfully completed a challenging thermography image segmentation project.
Interested in implementing a human in the loop in your AI pipeline? Get in touch with us and we would be happy to have a call.