Datasets:

• 800K food images and metadata from Instagram. [Health, Vision, Social Network Analytics]
• QMUL Drone Dataset – Images & Telemetry [ Vision, Domain Adaptation ]
• Shoe & Chair Data – Images & Sketches [ Fine-grained SBIR ]

Code:

• Sketch-A-Net: Deep Neural Network for Sketch Recognition
• MDMTL: Multi-task/multi-domain learning framework with neural networks.
• Sketch Me That Shoe: Code for CVPR’16 fine-grained SBIR paper.