eduzhai > Applied Sciences > Engineering >

Hand-drawn Symbol Recognition of Surgical Flowsheet Graphs with Deep Image Segmentation

  • Save

... pages left unread,continue reading

Document pages: 8 pages

Abstract: Perioperative data are essential to investigating the causes of adversesurgical outcomes. In some low to middle income countries, these data arecomputationally inaccessible due to a lack of digitization of surgicalflowsheets. In this paper, we present a deep image segmentation approach usinga U-Net architecture that can detect hand-drawn symbols on a flowsheet graph.The segmentation mask outputs are post-processed with techniques unique to eachsymbol to convert into numeric values. The U-Net method can detect, at theappropriate time intervals, the symbols for heart rate and blood pressure withover 99 percent accuracy. Over 95 percent of the predictions fall within anabsolute error of five when compared to the actual value. The deep learningmodel outperformed template matching even with a small size of annotated imagesavailable for the training set.

Please select stars to rate!

         

0 comments Sign in to leave a comment.

    Data loading, please wait...
×