In the novel paradigm, an image in the source domain is translated into the target domain, while the variations unrelated to the domain are maintained.

To address this issue, we propose a novel data augmentation paradigm that can adapt variations from one class to another.

First, current algorithms made variations only inside one specific class, but more desirable variations can further promote performance. But there are still two challenges: creating more desirable variations for scarce and imbalanced data, and designing a data augmentation to ease object detection and instance segmentation. Data augmentation, aiming to create variations for training data, has shown its power for this issue. To obtain a competing performance with a deep learning algorithm, enough amount of annotated data is requested but in the natural world, scarce or imbalanced data are common, and annotated data is expensive or hard to collect. 4College of Artificial Intelligence, Tianjin University of Science and Technology, Tianjin, Chinaĭeep learning shows its advantages and potentials in plant disease recognition and has witnessed a profound development in recent years.3Department of Computer Engineering, Mokpo National University, Jeonnam, South Korea.2Core Research Institute of Intelligent Robots, Jeonbuk National University, Jeonbuk, South Korea.1Department of Electronics Engineering, Jeonbuk National University, Jeonbuk, South Korea.

Mingle Xu 1,2, Sook Yoon 3 *, Alvaro Fuentes 1,2, Jucheng Yang 4 and Dong Sun Park 1,2 *

Picture translate