Person Image Generation To Enhance The Input Training Image For Re-Identification

The process of recognizing a person from photos or videos taken with a different camera is called re-identification (reID). In today’s world, the widespread use of extensive networks of CCTV and security cameras makes it difficult for a human operator to trace a person. This is where re-identification technology comes into play. However, person re-identification in real-world applications is very challenging due to the size of the existing dataset, the lack of human motions, and the inability to supply enough training samples. In this paper, proposed a method for person image generation using Generative Adversarial Networks (GANs). GANs allow us to build additional person images that look as realistic as possible.