장길진 교수(IT대학 전자공학부)

o (1997) B.S., Korea Advanced Institute of Science and Technology
o (1999) M.S., Korea Advanced Institute of Science and Technology
o (2004) Ph.D., Korea Advanced Institute of Science and Technology

o (2014~Present) Professor, Kyungpook National University, Daegu, Korea
o (2009~2014) Assistant Professor, Ulsan National Institute of Science and Technology, Ulsan, Korea
o (2006~2009) Postdoctoral Researcher, University of California, San Diego, USA
o (2004~2006) Research Staff, Samsung Advanced Institute of Technology, Giheung, Korea

Speech signal processing, audio signal processing, speech recognition, blind source separation, machine learning, medical signal processing, medical disease diagnosis

o Weakly Supervised U-Net with Limited Upsampling for Sound Event Detection
o Individualized diagnosis of preclinical Alzheimer’s Disease using deep neural networks
o Encoder-Weighted W-Net for Unsupervised Segmentation of Cervix Region in Colposcopy Images
o Hierarchical Phoneme Classification for Improved Speech Recognition
o Speech and music pitch trajectory classification using recurrent neural networks for monaural speech segregation
o Speaker Adaptation Using i-Vector Based Clustering
o Improving Singing Voice Separation Using Curriculum Learning on Recurrent Neural Networks
o Single channel blind source separation based on probabilistic matrix factorization

https://sites.google.com/site/milabknu/