Synthetic data is used in computer vision to develop and test algorithms that have to work on real data. The generation of realistic synthetic data has regained interest with the advent of deep-learning, proposing to opt for synthetically generated images rather than, or in addition to, real data in order to overcome the lack of suitable labelled large databases. A 2019 survey of the field calls the use of synthetic data "one of the most promising general techniques on the rise in modern deep learning, especially computer vision". We have been using it for human and machine learning process since 2003 within a Virtual Museum Simulator, but the idea of using artificially generated data to tune and validate algorithms, predates it by at least a decade: just think of the validation of OCR output applied to machine-printed documents generated specifically for the task.
HAND SHAPE RECOGNITION
We propose a novel approach to 3D hand shape recognition from RGB-D data based on geometric deep learning techniques. The model, trained on synthetic data, retains the performance on real samples during test time.
J. Svoboda, P. Astolfi, D. Boscaini, J. Masci and MM. Bronstein. Clustered Dynamic Graph CNN for Biometric 3D Hand Shape Recognition, IEEE International Joint Conference on Biometrics - IJCB, pp. 1-9, 2020
A key need in the development of algorithms in computer vision (as in many other fields) is the availability of large datasets for training and testing them. Ideally, datasets should cover the expected variability range of data and be supported by high quality annotations describing what they represent so that the response of an algorithm can be compared to reality. Gathering large, high quality data sets is however a time consuming effort. An alternative is available for computer vision research: computer graphics systems can be used to generate photo-realistic images of complex environments together with supporting ground truth information. This chapter shows how these systems can be exploited to generate a flexible (and cheap) evaluation environment.
R. Brunelli. Appendix B: Synthetic Oracles for Algorithm Development, in book: Template Matching Techniques in Computer Vision: Theory and Practice, Wiley, 2009
To study computer vision algorithms, specifically to develop low level pedestrian segmentation and tracking algorithms a virtual museum simulator, populated with scripted visitors. The virtual museum simulator uses 3D rendering techniques with support for global illumination, shadows and different visual artifacts such as motion blur and interlacing.
O. Lanz. Occlusion robust tracking of multiple objects, International Conference on Computer Vision and Graphics, 2004
A. Santuari, O. Lanz, and R. Brunelli. Synthetic Movies for Computer Vision Applications. IASTED International Conference: Visualization, Imaging, and Image Processing, pp. 1–6, 2003
F. Bertamini, R. Brunelli, O. Lanz, A. Roat, A. Santuari, F. Tobia and Q. Xu. Olympus: an ambient intelligence architecture on the verge of reality, International Conference on Image Analysis and Processing, pp. 139-144, 2003