Deep Learning for Logo Recognition

In this paper we propose a method for logo recognition using deep learning. Our recognition pipeline is composed of a logo region proposal followed by a Convolutional Neural Network (CNN) specifically trained for logo classification, even if they are not precisely localized. Experiments are carried out on the FlickrLogos-32 database, and we evaluate the effect on recognition performance of synthetic versus real data augmentation, and image pre-processing. Moreover, we systematically investigate the benefits of different training choices such as class-balancing, sample-weighting and explicit modeling the background class (i.e. no-logo regions). Experimental results confirm the feasibility of the proposed method, that outperforms the methods in the state of the art.

[1]  Radu Dogaru,et al.  Logo localization and recognition in natural images using homographic class graphs , 2015, Machine Vision and Applications.

[2]  Bernardete Ribeiro,et al.  Automatic graphic logo detection via Fast Region-based Convolutional Networks , 2016, 2016 International Joint Conference on Neural Networks (IJCNN).

[3]  Rainer Lienhart,et al.  Bundle min-hashing for logo recognition , 2013, ICMR '13.

[4]  Olivier Buisson,et al.  Logo retrieval with a contrario visual query expansion , 2009, ACM Multimedia.

[5]  Jihad El-Sana,et al.  Shape recognition and pose estimation for mobile augmented reality , 2009, ISMAR.

[6]  Raimondo Schettini,et al.  Logo Recognition Using CNN Features , 2015, ICIAP.

[7]  Alberto Del Bimbo,et al.  Trademark matching and retrieval in sports video databases , 2007, MIR '07.

[8]  Corneliu Florea,et al.  Elliptical ASIFT Agglomeration in Class Prototype for Logo Detection , 2015, BMVC.

[9]  Yann LeCun,et al.  Regularization of Neural Networks using DropConnect , 2013, ICML.

[10]  Jitendra Malik,et al.  Region-Based Convolutional Networks for Accurate Object Detection and Segmentation , 2016, IEEE Transactions on Pattern Analysis and Machine Intelligence.

[11]  Yue Gao,et al.  Brand Data Gathering From Live Social Media Streams , 2014, ICMR.

[12]  Geoffrey E. Hinton,et al.  ImageNet classification with deep convolutional neural networks , 2012, Commun. ACM.

[13]  Eero P. Simoncelli,et al.  Image quality assessment: from error visibility to structural similarity , 2004, IEEE Transactions on Image Processing.

[14]  Eleftherios Kayafas,et al.  Vehicle Logo Recognition Using a SIFT-Based Enhanced Matching Scheme , 2010, IEEE Transactions on Intelligent Transportation Systems.

[15]  Cordelia Schmid,et al.  Correlation-based burstiness for logo retrieval , 2012, ACM Multimedia.

[16]  Christian Eggert,et al.  On the Benefit of Synthetic Data for Company Logo Detection , 2015, ACM Multimedia.

[17]  Forrest N. Iandola,et al.  DeepLogo: Hitting Logo Recognition with the Deep Neural Network Hammer , 2015, ArXiv.

[18]  Kaiming He,et al.  Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks , 2015, IEEE Transactions on Pattern Analysis and Machine Intelligence.

[19]  Xing Xie,et al.  Spatial pyramid mining for logo detection in natural scenes , 2008, 2008 IEEE International Conference on Multimedia and Expo.

[20]  Rainer Lienhart,et al.  Scalable logo recognition in real-world images , 2011, ICMR.

[21]  Koen E. A. van de Sande,et al.  Selective Search for Object Recognition , 2013, International Journal of Computer Vision.

[22]  Yuning Jiang,et al.  Interactive visual object search through mutual information maximization , 2010, ACM Multimedia.

[23]  Koen E. A. van de Sande,et al.  Segmentation as selective search for object recognition , 2011, 2011 International Conference on Computer Vision.

[24]  Nitish Srivastava,et al.  Dropout: a simple way to prevent neural networks from overfitting , 2014, J. Mach. Learn. Res..