2016/04/20 by Gonçalo Oliveira, Oliveira, Gonçalo, Xavier Frazão +5
Computer Science · #Advanced Image and Video Retrieval Techniques #Computer Vision and Pattern Recognition (cs.CV) #FOS: Computer and information sciences #Image Retrieval and Classification Techniques #Multimodal Machine Learning Applications
paper · pdf · doi:10.48550/arxiv.1604.06083
openalex publication_date 2016/04/20 · openalex created_date 2022/10/04 · openalex updated_date 2026/07/28
Brand recognition is a very challenging topic with many useful applications\nin localization recognition, advertisement and marketing. In this paper we\npresent an automatic graphic logo detection system that robustly handles\nunconstrained imaging conditions. Our approach is based on Fast Region-based\nConvolutional Networks (FRCN) proposed by Ross Girshick, which have shown\nstate-of-the-art performance in several generic object recognition tasks\n(PASCAL Visual Object Classes challenges). In particular, we use two CNN models\npre-trained with the ILSVRC ImageNet dataset and we look at the selective\nsearch of windows `proposals' in the pre-processing stage and data augmentation\nto enhance the logo recognition rate. The novelty lies in the use of transfer\nlearning to leverage powerful Convolutional Neural Network models trained with\nlarge-scale datasets and repurpose them in the context of graphic logo\ndetection. Another benefit of this framework is that it allows for multiple\ndetections of graphic logos using regions that are likely to have an object.\nExperimental results with the FlickrLogos-32 dataset show not only the\npromising performance of our developed models with respect to noise and other\ntransformations a graphic logo can be subject to, but also its superiority over\nstate-of-the-art systems with hand-crafted models and features.\n