Research Interests
Computer vision, machine learning, information retrieval - metric learning, multi-modal representation learning, semi/self-supervised learning, object detection/recognition, search
Projects
Camera Search: Product detection and search for Amazon Shopping App mShop (2020--, Amazon)
T-VSE: Large-Scale Visual Semantic Embedding (2020, Amazon)
OSLD: Open-Set Deep Logo Detection (2019, Amazon)
Real-time vehicle and person detection/analytics on in-car and CCTV videos with deep networks (2018, NTU)
SDML: Soft deep metric learning for visual search (2018, NTU)
FashionFor: Fashion modeling and trend forecasting with deep networks (2018, NTU)
AutoMon: Automated car monitoring using audio-visual and thermal imaging (2017, NTU)
ProDuctS: Deep learning based product image search for e-commerce sites
Deep multi-task learning for visual recognition and search: Using deep convolutional neural networks for learning learning multiple tasks, such as scene, object, attribute, action recognition and retrieval
Mobile Eye Vision: Using computer vision and machine learning to build mobile applications to help the visually impaired; e.g., Turkish text reading, object/scene recognition (2016)
Smartscope: Smart electronic stethoscope for the acquisition and computer aided diagnosis of body sounds, in particular respiratory sounds (2016)
SICURA: Analysis of X-ray images for the automatic detection of illicit objects/materials for security applications (2010-2013, Technical Univ. of Kaiserslautern, Germany)
BilVideo-7: MPEG-7 compatible video indexing and retrieval system (2008-2010, Bilkent University, Dept. of Computer Engineering)