Survey on Videos Data Augmentation for Deep Learning Models

In most Computer Vision applications, Deep Learning models achieve state-of-the-art performances. One drawback of is the large amount data needed to train models. Unfortunately, in many are difficult or expensive collect. Data augmentation can alleviate problem, generating new from a smaller initial dataset. Geometric and color space image methods increase accuracy but often not enough. More advanced solutions Domain Randomization use simulation artificially generate missing data. algorithms usually specifically designed for single images. Most recently, have been applied analysis video sequences. The aim this paper perform an exhaustive study novel techniques point out future directions research on topic.