Dataset for "Adaptive Keyframe Selection for Indoor Positioning Using Controllable Kernel Density Estimation"

This dataset supports the article "Adaptive Keyframe Selection for Indoor Positioning Using Controllable Kernel Density Estimation". The dataset includes RGB image frames captured from various indoor environments, organized into folders by scene or location (e.g., corridors, staircases). Each folder contains frames extracted from video sequences and is optionally aligned with associated RSS (Wi-Fi signal strength) values. These data were used for evaluating the proposed CKDE framework for keyframe extraction that combines visual and RSS-based information to improve indoor positioning robustness.
The dataset may include:
Raw or preprocessed image framesOptional corresponding metadata (e.g., RSS values per frame or location)README file explaining folder structure, preprocessing steps, and formatThis dataset was collected at Ferdowsi University of Mashhad (Engineering Faculty) and is used in experiments described in Section 5 of the paper.