Liming Chen – From Deep Learning to Weakly Supervised Learning for Computer Vision Tasks
Deep learning has revolutionized a number of domains, including in particular the field of computer vision. However, it requires a large amount of labeled data for training. In this talk, I am presenting our recent work on several computer vision tasks, e.g., face recognition, edge detection, object detection, and show that we can design simple neural networks yet achieving amazing performance and in most cases even bypass the requirement of large labeled training data for an effective training, using synthesized data or transfer learning.
Liming Chen received the joint BSc degree in mathematics and computer science from the University of Nantes in 1984, the master’s degree in 1986, and the PhD degree in computer science from the University of Paris 6 in 1989. He first served as an associate professor at the Université de Technologie de Compiègne, then joined Ecole Centrale de Lyon (ECL) as a professor in 1998, where he leads an advanced research team in multimedia computing and pattern recognition. He has been the recipient of the prime of excellent research discerned by the Ministry of Research since 1995, and head of the Department of Mathematics and Computer Science at ECL between 2007 and 2016. His research work was the subject of 37 theses under Prof. Chen’s supervision or co-supervision, including 34 theses since his appointment to ECL, 295 publications, including 1 book, 11 book chapters, 45 international journals, 2 national journals, 154 international conferences, 39 international workshops, and 15 national conferences, filed 7 patents and 2 software, 14 national or European collaborative research projects for a total amount of € 2.8 million, 12 industrial contracts for a total amount of 777.6 K€ incl. Prof.Chen is on the board of the French Association for Recognition and Interpretation of Patterns (AFRIRF), the French branch of the IAPR, an associate editor for Eurasip Journal on Image and Video Processing, a guest editor for a special issue on “Situation, Activity and Goal Awareness in Cyber-physical Human-Machine Systems” for IEEE Transactions on Human-Machine Systems (THMS) and a senior member of the IEEE. His current research interest is computer vision and machine learning, including in particular 2D/3D face analysis and recognition, image categorization, object detection, and affective computing both in image, audio and video.