00000ctmaa22000004a 4500 UP-1685675941123990597 Buklod 20220603150211.0 a r |||| u| ta 180101s2018 xx a r |||| u|eng d (iLib)UPTAC-00020025776 UPVTC eng LG 993.5 2018 C66 A33 Abantao, Gerardo Jr. A. Pollen viability analysis using Watershed Segmentation and Support Vector Machine on selected rice genotypes (Oryza sativa L.) a special problem by Gerardo A. Abantao Jr.; Ryan Rey M. Daga, adviser. Tacloban City Gerardo A. Abantao Jr. 2018. xi, 77 leaves illustrations 29 cm. Undergraduate thesis (B.S. in Computer Science)--University of the Philippines Visayas Tacloban College, December 2018. Determining the number of instances of objects in an image is a common problem in computer vision. Solving this kind of problem manually requires a large amount of time. Pollen viability analysis is one example of object counting task. It needs a series of verification of outputs to have an accurate results. Pollen viability is determined by the percentage of pollen grains in an image. It plays a significant role for growers and breeders on monitoring the production of plants. Image processing and machine techniques was used for faster and efficient pollen viability analysis in this study. Watershed Segmentation and Support Vector Machines was used to pollen viability analysis of various genotypes of rice pollen. To measure the accuracy of the proposed approach, experiments was performed on two sets of stained pollen grains, the first set contains images without lumped and the second set contains pollen grains with lumped. Grid estimation was also used to improve the accuracy of classification of pollen grains on lumped region. In grid estimation the lumped region is divided into regions based on the average size of pollen in an image. Based on conducted experiments Watershed with SVM yielded a highest accuracy of 95.53% on set of pollen images without lumped and 91.30% on set of pollen images with lumped. The proposed approach shown to produced results faster than manual counting, which takes 1 to 5 seconds per image. Rice Biotechnology. Biology computing. Pollen viability analysis Watershed segmentations. Pollen viability analysis Support vector machines. Oryza sativa L. Daga, Ryan Rey M. adviser. FI UP UPTAC UPTAC LG 993.5 2018 C66 A33 Thesis