| Abstract | Dog owners usually have a special bond with their d ogs. It is undoubtedly devastating for them when their dog goes missing. Nowadays, there a re many ways to find lost dogs. The Internet, which has become a part of almost everyon e's life, is unquestionably one of the many waysto help spread the word. It is fast, wides pread, and does not require a lot of time or money. However, there are currently no websites in Thailand specifically developed to assist people in finding lost dogs and owners of fo und dogs. Therefore, this thesis aims to provide an option for people to find their lost dog s and owners of found dogs by developing a community website to serve as a centra l point where dog lovers can meet, share their experiences, support one another, and h elp return lost dogs to their homes, with assisting functionality including an easy-to-use do g reporting system, social network sharing, anotification system, a search system, and an automatic dog face recognition system. To identify individual dogs, I adopt local binary patterns (LBP). 1,513full-frontal pictures of dogs of 119 different breeds were colle cted for the experiment, and the LBP technique achieved an accuracy of 86.25%. It worked well with pictures under a wide range of illumination conditions but was not robust to pictures acquired under extreme illumination conditions, for example, dark shadows. I evaluated the Website by having 20 people use it, and they all agreed that they were s atisfied with the Website and would use it when it was officially launched. |