Estimating Bicycle Route Attractivity from Image Data
This work addresses a domain-specific problem for urban planning or cycling communities, but it is incremental as it applies existing methods to a new dataset.
The thesis tackled the problem of scoring bicycle route attractivity using image data by applying Convolutional Neural Networks to a dataset enhanced with Google Street View and Open Street Map information, resulting in experiments with various techniques and architectures to improve scoring.
This master thesis focuses on practical application of Convolutional Neural Network models on the task of road labeling with bike attractivity score. We start with an abstraction of real world locations into nodes and scored edges in partially annotated dataset. We enhance information available about each edge with photographic data from Google Street View service and with additional neighborhood information from Open Street Map database. We teach a model on this enhanced dataset and experiment with ImageNet Large Scale Visual Recognition Competition. We try different dataset enhancing techniques as well as various model architectures to improve road scoring. We also make use of transfer learning to use features from a task with rich dataset of ImageNet into our task with smaller number of images, to prevent model overfitting.