In this thesis, we study an open set recognition algorithm that is based on the Sparse Representation-based Classification (SRC) method. By modeling the tail distributions of the matched and non-matched reconstruction errors using the statistical Extreme Value Theory (EVT), we simplify the open set recognition problem into a set of hypothesis testing problems. The confidence scores corresponding to the tail distributions of a novel test sample are then fused to determine its identity. The

more »

... iveness of the proposed method is demonstrated using three publicly available image and object classification datasets and it is shown that this method can perform significantly better than many competitive open set recognition algorithms. ii Acknowledgements The whole academic life is somehow like training a model, which adjusts its 'parameter' everyday to fit in the new sample and increase the 'accuracy '. I am lucky that I achieve the 'accuracy' that can make it to the defense with the help of so many talented people. Firstly, I'd like thank my advisor Prof. Vishal M Patel for guiding me and supporting me through my second year of my master degree. He is more like a brother that give me a lot of help both on research and life. I am looking forward to pursuing Phd degree under his advise.