00000ctmaa22000003a 4500 UP-1685675941123958235 Buklod 20180817122159.0 aa rb |||1 u| ta 180817s xx d r |||| u| (iLib)UPTAC-00019010355 UPVTC eng LG 993.5 2017 C66 P35 Paloma, Anna Rose B. SamPoll an application for ballot mark classification using a sample set of Operator-labeled Vote Marks. Anna Rose B. Paloma 2017 77 leaves illustrations Undergraduate thesis (B.S. Computer Science)-- University of the Philippines, Tacloban. We present SamPoll: an application that classifies ballot vote marks using a sample set of operator-labeled vote marks. SamPoll is designed to help election auditors classify different types of vote marks, other than the usual filled-in ovals, as valid or invalid during post-election audits. Using scanned ballot images as input, the application performs a series of image processing techniques to accurately extract every vote marks on the ballots. With the use of feature extraction, feature vectors that distinguish every note marks are fed into a machine learning algorithm to create a classifier. We tested our application on scanned ballot templates that were used during the 2016 Philippine national election. The results of this study have shown that the application was able to classify vote marks with high accuracy, especially in recognizing ideally filled-in ovals. FI UP UPTAC UPTAC LG 993.5 2017 C66 P35 Thesis