July 2016: From Perceptron to Deep ConvNets: Achievements & Challenges in Visual Applications
On 14th of July 2016, Dr. Andrew Teoh (associate professor from Yonsei University, Korea) visited Centre for Cyber Security for his research sharing and discussions with the member of CCS. During the visit, Dr. Andrew Teoh gave a presentation on “From Perceptron to Deep ConvNets: Achievements & Challenges in Visual Application”. In the talk, Dr. Andrew focused on convolution networks (ConvNet), which was shown enormously successful in visual related applications. He gave a historical overview of neural networks and Convnet research by tracing back to four decades ago. The fundamental ingredients as well as the strengths and weaknesses of Convnet were discussed. The reasons behind the recent success of Convnet and its high impact results were highlighted too.
Besides, Dr. Andrew Teoh met some of CCS postgraduate students to discuss their research on biometric recognition. Recently, the discussion result on “Cancellable Iris Template Generation based on Indexing-First-One Hashing” (a joint work between Yen-Lung Lai, Zhe Jin, Andrew Beng-Jin Teoh, Bok-Min Goi, Wun-She Yap, Tong-Yuen Chai and Christian Rathgeb) had been published in Pattern Recognition.