Project JaundiceAI

Jaundice index inference: our derived (from RGB) 13-band ML model during tryout has reached R² = 83.8%

(Pearson correlation coefficient=0.92, with RMSE=0.3782) for 90 patient images.



Training and Validation Loss over 285 Epochs




Prediction performance with R²=83.8% (PCC=0.92) using JaundiceAI tryout model (need more data)




Below are samples from 141 HSI bands: (we are fundraising to collect more HSI data)



AITek is raising $1M fund to work with healthcare partners and instrument manufacturer to collect more comprehensive data and test HSI/mobile App.s

Worldwide Markets

With strong hyperspectrum patents and AI/ML applications, we are posing to address to the following markets:

• In Vitro Diagnostics (IVD, replace blood test with real-time generative jaundice index)
• Skin Analysis/Diagnostics
• Food Safety/Chemicals Testing
• Alzheimer's and Parkinson’s Disease early detection
• Pancreatic cancer early detection
• Global Vision Screening and Diagnostics

Contact Us

Santa Clara, CA, USA