THE PROJECT
Spek Diagnostics
Spek Diagnostics is a Garden City, Idaho dermatopathology lab. We built a custom whole-slide viewer pathologists can actually diagnose from, and we are adding AI image analysis with HIBOU and PathoSAM so the software can help flag and segment tissue, not only pan and zoom.
Mission
Spek Diagnostics provides prompt, accurate pathology results to dermatologists and other specialists across the Treasure Valley and beyond. Their pathologists work from digital whole-slide images, not just a microscope. They needed a viewer that matched how they actually read slides, and they wanted AI on those images that could aid diagnosis without pretending to replace the pathologist.
Tech
The hard problems are streaming tiled slide data, keeping pan and zoom feeling instant, applying scanner color profiles, and running pathology models against those images. Python and ML (HIBOU, PathoSAM) sit next to OpenSlide for reading whole-slide files and OpenSeadragon for deep-zoom viewing in the browser. Docker keeps that pipeline portable.
Python / ML
Python is how we do serious data and model work. The scientific stack, notebooks-to-services path, and the ecosystem around imaging and machine learning mean we can move from a research model to something a product can call without switching languages halfway.
Python is the language for the analysis side: preprocessing slides, running HIBOU for histopathology representations, and PathoSAM for segmentation, so the viewer can aid diagnosis instead of only displaying pixels.
EMBARK ON THIS EXTRAORDINARY
JOURNEY WITH US
Are you ready to see your vision come to life? We would love to hear from you!