1 / 10

REU Summer 2002 Weekly Report 3

REU Summer 2002 Weekly Report 3. Lorraine Harrington University of Nevada, Reno. Research Goals. Develop a medication monitoring program for the elderly Have patient/medical staff build a database of medications (times & descriptions)

sakina
Download Presentation

REU Summer 2002 Weekly Report 3

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. REU Summer 2002Weekly Report 3 Lorraine Harrington University of Nevada, Reno

  2. Research Goals • Develop a medication monitoring program for the elderly • Have patient/medical staff build a database of medications (times & descriptions) • Use a two-camera system hooked into a personal computer to monitor the medication consumption • Check the meds taken against the database – set off alarm if a medication is missed or the wrong medication is taken

  3. Overview • Focus on medication consumption • Currently, assume that if: • Pill bottle is picked up and opened • Mouth opens • The hand occludes the mouth while it is open • Swallowing • Then the pill was taken

  4. Problem 1: Identify Pill Bottles

  5. Pill Bottle Identification • Background subtraction • Connected components • Double size thresholding • If pixel remains on, look at orig. image (image addition)

  6. Problems • Salt shaker is identified as a bottle • The three pill bottles have heads, necks, and trunks. The shaker tapers from the head. Do edge detection to find the shape of the head – if cylinder or has neck, is a pill bottle.

  7. When arm moves, bottle & arm are seen as one component and deleted by threshold • Initialize the pill bottles with a stationary person, then track the pill bottles to maintain definition even when in contact/occluded with another object.

  8. Differentiate between prescription bottles and OTC bottles • Currently, just identifies bottles • After running ConnectedComponents to identify bottles, use the linked list of each pixel in the object to check the color of the original pixel • If the color is within a preset range (yellow-orange values) increment a counter • If the final counter value for the object is above a threshold (50 pixels), mark the object as a prescription – if not, as an OTC drug.

  9. Next Week • Implement problem solutions • Use skin detection to find user’s head and arms • Use component tracking to follow bottle when picked up by user • Detect twisting of hand as unscrewing the cap on a pill bottle

  10. Research • Methods of mouth detection • Need to track mouth opening/closing • Handling mouth/hand occlusion – measure distance between hand and mouth • Detecting swallowing in females

More Related