Image Data Visualization Using T-SNE for Urban Pavement Disease Recognition

Abstract Urban pavement disease recognition is for the most part, a mission performed manually. Recently, video analysis task has been one of important applications in various fields. Aims to renovate on automated vision-based and experiment new methods road detection, this paper analyzes images took from city image data visualization issues. From features extracted histogram oriented gradient, we perform principal components reduce dimensionality features. Results are presented by visualizing datapoints using t-distributed stochastic neighbor embedding. The shows that t-SNE suitable growing field urban recognition.