Pr. Siham Tabik
Pr. Siham Tabik

Affiliation

Dept.  Computer Sciences and Artificial Intelligence

Research group "Soft Computing and Intelligent Information Systems"

University of Granada

 

 

Tutorial: Deep Convolutional Neural Networks for image classification and object detection

Abstract

In the last five years, deep learning in general and Convolutional Neural Networks (CNNs) in particular have shown impressive accuracies compared to all the classical methods in pattern recognition, classification and detection especially in the field of computer vision. In fact, since 2012, the prestigious Large Scale Visual Recognition Challenge (ILSVRC) has been won only by CNNs. CNNs are showing a high potential for learning more and more complex patterns. They have been used with success in diverse problems, e.g., image classification, handwritten digit recognition and speech recognition, and they are being explored in more fields and applications. This tutorial will provide the state-of-the art of CNNs-based models for object classification and detection, give an overview of practical techniques that improve the prformance of CNNs-based models, namely, data- augumentation, fine-tuning, transfer learning, then present some successful case studies, e.g.,  hand-written digit classification problem, known as MNIST and an automatic handgun detection model in videos.

Biography

Siham Tabik received the B.Sc. degree in physics from University Mohammed V, Rabat, Morocco, in 1998. From 2002 to 2006, she carried out her PhD studies in computer science at the University of Almeria. From 2007 to 2015 she worked as Postdoc researcher funded by three competitive Postdoc programs at the University of Malaga, Spain. From 2015 to 2016 she served as assistant professor at Dpt. of Computer Science and Systems Engineering, University of Zaragoza, Spain. Since Oct. 2016  Siham is a Ramón y Cajal researcher at the University of Granada, Spain and also member of the Soft Computing and Intelligent Information Systems research group.

Her current research interests include the design of scalable algorithms, High Performance Computing (HPC) and deep learning models. She published over 34 articles in indexed journals and Conferences.

She was awarded the 2009 Eurocities Innovation award for the interactive solar energy website. She got the Spin-off prize from the chamber of Commerce, in Málaga, Spain, in 2009 to create the company of Helioserver. She also received the 2010 Itanium Innovation Award for the most efficient intensive application implementation. Recently, she received the first prize for the best innovative project in security given by Security Forum.

She also was Postdoctoral Visiting Research Associate at 1) Dpt. of Computer Science, University of Virginia, USA, during six months in 2007, 2) Laboratory of Regional analysis and Remote Sensing, University of Buenos Aires, Argentina, during 3 months in 2008, 3) Dpt. of Computer Science, University of Austin, USA, during six months in 2009, 4) Dpt. of Electrical Engineering, Eindhoven University of Technology, Netherlands, during five months in 2015.