The goal of the ISL Image Databases is to share image data sets with researchers around the world. The ISL image databases consist of five major parts: medical image databases, human fatigue image databases, face and eye databases, facial expression databases, and human activity databases.
This database can be used in many applications such as face detection, eye detection, and face pose estimation.
This database consists of video sequences and static images. It can be used in many applications such as face and eye detection, face tracking, face pose estimation, and face recognition.
This database consists of image sequences taken under IR cameras. Each image sequence is decomposed into dark pupil image sequence and its corresponding bright pupil image sequence. This database is primarily used in eye detection and tracking.
The training image databases contain the cropped image region of interest (ROI) for training various classifiers including eye, eye corner, and face.
This database can be primarily used for facial action unit recognition. It consists of two subsets: frontal-view facial expression database and multi-view facial expression database.
Our research group has manually created more groundtruth data for Cohn-Kanade AU-coded facial expression database including the frame-by-frame AU labeling and coordinates of 34 facial feature points.
The RPI ISL databases are publicly available for non-commercial use. Any commercial distribution or any act related to commercial use of this database is strictly prohibited. You can follow the following steps to get access to RPI ISL databases:
Each database includes several sub-databases for different use. Please read the descriptions carefully and select the databases you may be interested. Please note that only the databases with license are available.
If you are working on a research project, please ask the Principal Investigator of the project to sign the license agreement form, which you can download through the link beside each database.
Please sign and scan the license and send it to Professor Qiang Ji (qji@ecse.rpi.edu) vie either email or fax.
For graduate students who are interested in the data, please ask your research advisor to sign the license for you. Please note that we will silently ignore the inquiry if the license is signed by the student. And only the databases with license are available.