This is the synthetic data set used in the paper 'Combining Deep Learning with Geometric Features for Image based Localization in the Gastrointestinal Tract' published in Expert Systems with Applications.

The two sequences are 00 and 01. 00 is for training, and 01 is for testing.

The camera pose file 0X.txt is in this format: [Filename] [position X] [position Y] [position Z] [Quaternion W] [Quaternion X] [Quaternion Y] [Quaternion Z]

The images strictly follow the pinhole model. All images have no distortion. The intrinsic matrix is: 520 0 320 0 520 240 0 0 1

Readers are free to use the data set for any purpose. It comes with no warranty! For academy use, please consider citing our paper:

@article{song2020combining, title={Combining Deep Learning with Geometric Features for Image based Localization in the Gastrointestinal Tract}, author={Song, Jingwei and Patel, Mitesh and Girgensohn, Andreas and Kim, Chelhwon}, journal={Expert Systems with Applications}, year={2021} }

Please contact Jingwei Song ([email protected]) if you have any enquery.