Border Ownership and Category Annotation extended from Virtual Kitti 2

This dataset provides extended annotations for the Virtual KITTI 2 (VKITTI2) dataset to support research on border ownership, disparity, and category selectivity-based segmentation, as well as figure-ground organization in synthetic driving scenes. It includes: 2-channel border ownership maps for object contours (Chen et al., 2022) 1-channel (vehicle) category channel maps (Chen et al., 2022) Near and Far Absolute disparity maps (full and downscaled resolution) (Chen et al., 2025) MATLAB and Python scripts for reading, writing, and visualizing annotations in `.flo` format These annotations were developed as part of a computational framework for figure-ground segmentation inspired by visual cortex processing. The dataset is designed to be used alongside the original Virtual KITTI 2 dataset (Cabon et al., 2020), which must be downloaded separately from NAVER LABS Europe. **License:** This dataset is released under the Creative Commons Attribution-Non-Commercial-ShareAlike 3.0 (CC-BY-SA 3.0) license. It inherits the ShareAlike clause due to its use of derivative annotations based on VKITTI2. ('LICENSE.txt' is also included in the tar zipped package) **How to Use:**- (Optionally) Download the original VKITTI2 RGB and depth data from: https://europe.naverlabs.com/research/computer-vision/proxy-virtual-worlds-vkitti-2/- Reference the included `README.md` and `DATAFORMAT.md` for format specifications and usage instructions