The PS_Sculpture training dataset introduced by the PS-FCN [1] contains various non-Lambertian reflectances, cast shadows, interreflections and effective noise information. However, for dark materials such as black-phenolic and steel, significant data loss happens due to 8-bit quantification. To lessen this data loss, we design a new supplementary training dataset rendered by 10 blobby objects and 10 other objects freely downloaded from the Internet and the real BRDF data comes from the MERL dataset [2].


These simulated live cell microscopy sequences were generated by the CytoPacq web service [R1]. The dataset is composed of 51 2D sequences and 41 3D sequences. The 2D sequences are divided into distinct 44 training and 7 test sets. The 3D sequences are divided into distinct 34 training and 7 test sets. Each sequence contains up to 200 frames.


Data augmentation is commonly used to increase the size and diversity of the datasets in machine learning. It is of particular importance to evaluate the robustness of the existing machine learning methods. With progress in geometrical and 3D machine learning, many methods exist to augment a 3D object, from the generation of random orientations to exploring different perspectives of an object. In high-precision applications, the machine learning model must be robust with respect to the small perturbations of the input object.


Datasets for image and video aesthetics

1. Video Dataset : 107 videos This dataset has videos that can be framed into images.

Color contrast,Depth of Field[DoF],Rule of Third[RoT] attributes

that affect aesthetics can be extracted from the video datasets.


2.Slow videos and Fast videos can be assessed for motion

affecting aesthetics




This dataset contains about 140,000 Tweets related to exoskeletons. that were mined for a period of 5-years from May 21, 2017, to May 21, 2022. The tweets contain diverse forms of communications and conversations which communicate user interests, user perspectives, public opinion, reviews, feedback, suggestions, etc., related to exoskeletons.


The dataset contains only tweet identifiers (Tweet IDs) due to the terms and conditions of Twitter to re-distribute Twitter data ONLY for research purposes. They need to be hydrated to be used. The process of retrieving a tweet's complete information (such as the text of the tweet, username, user ID, date and time, etc.) using its ID is known as the hydration of a tweet ID. For hydrating this dataset the Hydrator application (link to download and a step-by-step tutorial on how to use Hydrator) may be used.


Data Description

This dataset consists of 7 .txt files. The following shows the number of Tweet IDs and the date range (of the associated tweets) in each of these files. 

Filename: Exoskeleton_TweetIDs_Set1.txt

Number of Tweet IDs – 22945, Date Range of Tweets - July 20, 2021 – May 21, 2022

Filename: Exoskeleton_TweetIDs_Set2.txt

Number of Tweet IDs – 19416, Date Range of Tweets - Dec 1, 2020 – July 19, 2021

Filename: Exoskeleton_TweetIDs_Set3.txt

Number of Tweet IDs – 16673, Date Range of Tweets - April 29, 2020 - Nov 30, 2020

Filename: Exoskeleton_TweetIDs_Set4.txt

Number of Tweet IDs – 16208, Date Range of Tweets - Oct 5, 2019 - Apr 28, 2020

Filename: Exoskeleton_TweetIDs_Set5.txt

Number of Tweet IDs – 17983, Date Range of Tweets - Feb 13, 2019 - Oct 4, 2019

Filename: Exoskeleton_TweetIDs_Set6.txt

Number of Tweet IDs – 34009, Date Range of Tweets - Nov 9, 2017 - Feb 12, 2019

Filename: Exoskeleton_TweetIDs_Set7.txt

Number of Tweet IDs – 11351, Date Range of Tweets - May 21, 2017 - Nov 8, 2017


For any questions related to the dataset, please contact Nirmalya Thakur at


Skeleton datasets for Normal, Antalgic, Stiff legged, Lurching, Steppage, and Trendelenburg gaits.


Sequential skeleton and average foot pressure data for normal and five pathological gaits (i.e., antalgic, lurching, steppage, stiff-legged, and Trendelenburg) were simultaneously collected. The skeleton data were collected by using Azure Kinect (Microsoft Corp. Redmond, WA, USA). The average foot pressure data were collected by GW1100 (GHIWell, Korea). 12 healthy subjects participated in data collection. They simulated the pathological gaits under strict supervision. A total of 1,440 data instances (12 people x 6 gait types x 20 walkings) were collected.


This is a unique energy-aware navigation dataset collected at the Canadian Space Agency’s Mars Emulation Terrain (MET) in Saint-Hubert, Quebec, Canada. It consists of raw and post-processed sensor measurements collected by our rover in addition to georeferenced aerial maps of the MET (colour mosaic, elevation model, slope and aspect maps). The data are available for download in human-readable format and rosbag (.bag) format. Python data fetching and plotting scripts and ROS-based visualization tools are also provided.


The entire dataset is separated into six different runs, each covering different sections of the MET at different times. The data was collected on September 4, 2018 between 17:00 and 19:00 (Eastern Daylight Time). The data is available in both human-readable format and in rosbag (.bag) format.

To avoid extremely large files, the rosbag data of every run was broken down into two parts: “runX_clouds_only.bag” and “runX_base.bag”. The former only contains the point clouds generated from the omnidirectional camera raw images after data collection, and the latter contains all the raw data and the remainder of the post-processed data. Both rosbags possess consistent timestamps and can be merged together using bagedit for example. A similar breakdown was followed for the human-readable data.

Aside from point clouds, the post-processed data of every run includes a blended cylindrical panorama made from the omnidirectional sensor images, planar rover velocity estimates from wheel encoder data and an estimated global trajectory obtained by fusing GPS and stereo imagery coming from cameras 0 and 1 of the omnidirectional sensor using VINS-Fusion later combined with the raw IMU data. Global sun vectors and relative ones (with respect to the rover’s base frame) were also calculated using the Pysolar library. This library also provided clear-sky direct irradiance estimates along every pyranometer measurement collected. Lastly, the set of georeferenced aerial maps, the transforms between different rover and sensor frames, and the intrinsic parameters of each camera are also available.

We strongly recommend interested users to visit the project's home page, which provides additional information about each run (such as their physical length and duration). All download links on the home page were updated to pull from the IEEE DataPort servers. A more detailed description of the test environment and hardware configuration are provided in the project's official journal publication.

Once the data products of the desired run are downloaded, the project's Github repository provides a lightweight ROS package and python utilities to fetch the desired data streams from the rosbags.


We propose a novel high-resolution dataset named, “Dataset for Indian Road Scenarios (DIRS21)” for developing perception systems for advanced driver assistance systems.


Opportunity++ is a precisely annotated dataset designed to support AI and machine learning research focused on the multimodal perception and learning of human activities (e.g. short actions, gestures, modes of locomotion, higher-level behavior).