Visual object recognition using deep convolutional neural network

This thesis report is submitted in partial fulfillment of the requirements for the degree of Bachelor of Science in Computer Science and Engineering, 2017.Cataloged from PDF version of thesis report.Includes bibliographical references (page 30-31).Visual object recognition has been lying at the convergence point between machine learning, computer vision and AI since the very beginning. From robotics to information retrieval, many desired applications demand the ability to identify and localize objects into different categories. Despite a number of object recognition algorithms and systems being proposed for a long time in order to address this problem, there still lacks a general and comprehensive solution for the modern challenges. Most prominently, new approaches and computational models of vision to analyzing data, such as the convolutional neural networks (CNNs), have enabled a much more nuanced understanding of visual representation. In this paper, I have proposed a deep CNN model to solve the aforementioned problem of object recognition and reported a promising performance on a benchmark classification dataset called CIFAR10.Sabbir AhmedB. Computer Science and Engineerin