DEED - Deep Digital
Description
The DEED project aims to develop a system based on deep learning techniques for digital marketing. Deep learning has gained major popularity in automated feature extraction from images, audio and text. The project tackles in a first phase a graphic logo detection system based on a fast region-based convolutional networks (FRCN) and transfer learning. In a second case phase a robust facial emotions recognition based on an improved version of the classic CNN-LeNet-5 will be built and compared to deeper networks such as GoogleNet and AlexNet. Last but not least, the system will look at recognition of brands and emotions in video media for digital advertising and marketing.Researchers
Bernardete Ribeiro (coordinator)
Catarina Silva
Joel P. Arrais
André Pimentel
Gonçalo Oliveira
Ana Filipa Folgado Laranjeira
Catarina Silva
Joel P. Arrais
André Pimentel
Gonçalo Oliveira
Ana Filipa Folgado Laranjeira
Partners
EyeSeeTotal budget
20 000,00 €Local budget
20 000,00 €Keywords
brand detection; emotion recognition; deep learning; CNN; advertising; digital marketingStart Date
2016-09-01End Date
2018-09-01Journal Articles
Conference Articles
2016
(3 publications)- Oliveira, G. and Pimentel, A. and Ribeiro, B. , "Autonomic Graphic Logo Detection via Fast Region-based Convolutional Networks", in IEEE International Joint Conference on Neural Networks (IJCNN), Vancouver, Canada, July 24-29, 2016
- Laranjeira, A.F.F. and Pimentel, A. and Ribeiro, B. , " How Deep can we Rely on Emotion Recognition", in IberoAmerican Congress on Pattern Recognition, LNCS, Springer, 2016
- Laranjeira, A.F.F. and Frazão, X. and Pimentel, A. and Ribeiro, B. , "A simple Net for a Deep Problem - Emotion Recognition", in Portuguese Conference on Pattern Recognition, 2016