This repository contains the datasets, raw and otherwise used in our (Ashutosh Tiwari, Aditya Patwardhan) CSE 570 Wireless Systems and Networking Course Project at Stony Brook University. This course was taught by Professor Shubham Jain. The course website (if online) can be found here.
The data is contained in the Zip Files. Since there is approx 300GB of data (in the compressed format), it has been uploaded to Google Drive.
You can find the data here.
(https://drive.google.com/drive/folders/1fq2zsYxU7wuyFZo_Cs-kfliL1KTGYX_4?usp=sharing if the embedded URL does not work.)
The content is as follows.
- AllRawImage.zip - All the raw image (png) captured for scene0 in the ViFit dataset.
- RanAll.zip - The synchronised data from a project that uses the ViFit data (referenced below)
- RAN4model_dfv4p4.zip - The synchronised data from a project that uses the ViFit data (referenced below) that was used to train a specific model.
- SingleTestSceneData_UnSynced.zip - As the name suggests, the unsynchronised camera and sensor data from a singular data collection iteration.
- ResultsData.zip - A zip file containing all the data our project generated (i.e. annotated images, combined sensor readings, etc.)
- Our project was inspired by the following paper:
Zhang, Haichao, et al. "OOSTraj: Out-of-Sight Trajectory Prediction With Vision-Positioning Denoising." Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. 2024. (pdf) - The raw Vifit datasets were acquired from the ViFit webpage.
- The cleaned, timestamp synchronised datasets (image and sensor) were acquired from another project using the ViFit dataset (thanks Bryan Bo Cao!!)