DoR

Scopus Indexed Publications

Title: Automatic Detection and Recognition of Object to Help Visually Impaired People while Visiting Liberation War Museum in Bangladesh

Author: Md. Mehedi Hasan, Afsana Akther Ankhi, Khairul Alam, Md. Nahiduzzaman Sajeeb, Md. Rejaul Alam, Rubaiya Hafiz,

Abstract: This paper represents an analysis of object detection by multi-image classification and machine learning techniques for blind people. Our system has been developed based on Speech Recognition that helps blind people to know the information about the different components in the liberation war museum in Bangladesh. From the initial survey it concerned that most of the blind people are excited to know what is inside in the liberation war museum, but there is no automated system in the museum to recognize the object as an explanation for them. For this thoughtfulness, we built a model for visual imperial people that will detect automatically and that provide an explanation about the object by speech recognition technique. For the whole process, we used Different Machine Learning(ML) tools like Scikit-learn, Pandas, Matploatlib, Numpy, TensorFlow. For preprocessing images TensorFlow, for machine learning we have used Scikit-learn. For measuring the accuracy of our work we used five different Machine Learning(ML) algorithms K-Nearest Neighbor(KNN), Support Vector Machine(SVM), Decision Tree Classifier, Random Forest, Naive Bayes and most popular image processing algorithm Convolutional Neural Network(CNN) to find out which algorithm gave the highest accuracy. Finally, we detect museum spectacles by this algorithm which produced the highest accuracy.

Keywords: Image Classification , Machine Learning , Deep Learning , Object Detection , Prediction , SVM , CNN

Journal or Conference Name: 2021 5th International Conference on Electrical Engineering and Information Communication Technology (ICEEICT)