Dataset: Variable Message Signal Annotated Images for Object Detection

This publication presents a dataset consisting of Spanish road images taken from inside vehicle, as well annotations in XML files PASCAL VOC format that indicate the location Variable Message Signals within them. Additionally, CSV file is attached with information regarding geographic position, folder where image located and text Spanish. can be used to train supervised learning computer vision algorithms such convolutional neural networks. Throughout this work, process followed obtain dataset, acquisition labeling its specifications are detailed. The constitutes 1216 instances, 888 positives 328 negatives, 1152 jpg resolution 1280 × 720 pixels. These divided into 756 real created data-augmentation technique. purpose help research since there not one specifically for VMSs.