Prominent Attribute Modification using Attribute Dependent Generative Adversarial Network

Modifying the facial images with desired attributes is important, though\nchallenging tasks in computer vision, where it aims to modify single or\nmultiple attributes of the face image. Some of the existing methods are either\nbased on attribute independent approaches where the modification is done in the\nlatent representation or attribute dependent approaches. The attribute\nindependent methods are limited in performance as they require the desired\npaired data for changing the desired attributes. Secondly, the attribute\nindependent constraint may result in the loss of information and, hence, fail\nin generating the required attributes in the face image. In contrast, the\nattribute dependent approaches are effective as these approaches are capable of\nmodifying the required features along with preserving the information in the\ngiven image. However, attribute dependent approaches are sensitive and require\na careful model design in generating high-quality results. To address this\nproblem, we propose an attribute dependent face modification approach. The\nproposed approach is based on two generators and two discriminators that\nutilize the binary as well as the real representation of the attributes and, in\nreturn, generate high-quality attribute modification results. Experiments on\nthe CelebA dataset show that our method effectively performs the multiple\nattribute editing with preserving other facial details intactly.\n