Geoinnovation+

Project description

Objective of the project was to deliver a trained ML model that could detect the following road damages: cracks, potholes, alligator cracks, side road damages. Overall project scope included: design of the multiuser annotation production pipeline, data annotation using customer’s data, training, validation, testing and benchmarking data preparation, training data extraction utilities, training data check,model hyperparameter tuning, model training, model evaluation, model benchmarking.