featured projects:   CovidID  Seeing Eye Drone    blindBike    SCD    iSight   iVest   ULearn   USquat

 

iSight

 

 

 

 

Uses Deep Learning to perform the task of Object Detection and the results are then communicated to the user using one of three visualization modules: Icon, Enhance and Cartoon. The icon mode uses a icon image that it magnified for easy viewing. The enhance performs image balancing and interpolated magnification of the original image ROI corresponding to the detected object's bounding box. Finally, the cartoon image merges the results of a algorithm featuring bilateral filtering of the image ROI and a second algorithm featuring adaptive thresholding on the image ROI into create a cartooned image that emphasizes edge imformation at the same time as iconifying the color informaiton.