Deres, Samantha Amor P. creator author. Munar, Caroline T. author Prades, Dave M. author Saunar, Zyndrek John Paul D. author Tabuena, Brianna Carlota L. author Oñate, Joseph Jessie, adviser Onesa, Rosel O., panelist, chair Fortuno, Kaela Marie N., panelist Ibo, Allan O. Jr., panelist text bibliography ph monographic eng

print

104 leaves. These days, an increasing number of tasks that require image recognition use convolutional neural networks. This research suggested utilizing a convolutional approach to implement road sign recognition and it focused on dealing with uploading an image and quickly identifying which kind of road sign is present in which class of an input image. Generally, this study aimed to develop a web application that uses a Convolutional Neural Network (CNN) algorithm to recognize road sign images. Additionally, a simple and easy prototype for recognizing and classifying road signs was created using CNN algorithms in order to achieve more accurate recognition results. After the training and testing process, the trained CNN model achieved 80% performance. The model's evaluation in continous testing and training, which shows that CNN is capable of recognizing images, shows that it is successful in predicting and recognizing road signs. This evaluation was achieved using the sklearn and keras modules of Python that provides a Python interface for CNN models. Furthermore, the proposed prototype achieved good performance for road sign recognition, which provides a better recognition performance hence, according to the experimental results on the Convolutional Neural Network and according to the generated average analysis, it is suitable for image recognition. Samantha Amor P. Deres, [and 4 others]. Undergraduate Theses (BSCS) - Computer Science Camarines Sur Polytechnic Colleges, 2022. Includes bibliographical references. CSPC 240117 20240117175641.0 eng