TY - JOUR AU - MC Supriya, Kiran Kumar JP, PY - 2020/03/20 Y2 - 2024/03/29 TI - TOWARDS REAL TIME LOGO DETECTION AND CLASSIFICATION USING DEEP LEARNING JF - International Journal of Control and Automation JA - IJCA VL - 13 IS - 02 SE - Articles DO - UR - http://sersc.org/journals/index.php/IJCA/article/view/6413 SP - 63 - 73 AB - In the recent years, there are significant changes in the methods used to solve the problems associated with logo detection, localization and classification. Deep learning has helped to achieve state of art results for logo detection, localization and classification. Key challenges with respect to logo detection network are to identify the location of the logo. Essentially the algorithms related to region proposal methods play a key role in categorizing the efficiency of logo detection networks. Some of the methods like SPPnet [1] and Fast RCNN [2] are efficient in bringing down the time required for detection, but it also indicates the overhead in terms of computation required for region proposal. This paper discusses about region proposal network (RPN) which enables and improves the efficiency of logo region proposals at real time. Our proposal includes method to extract the objects from content, feature selection and classification of the extracted object using convolutional neural network (CNN). Hence the proposed method consists of recognition pipeline which takes care of region proposal, logo classification based on specifically trained convolutional neural network. ER -