Sconti Face Style Transformation

Sconti Face Style Transformation

Gunashree K L1, Vidya S2

1 Student, Department of MCA, Bangalore Institute of Technology, Karnataka, India

gunashreekl12@gmail.com

2Professor, Department of MCA, Bangalore Institute of Technology, Karnataka, India

vidyas@bit-bangalore.edu.in

ABSTRACT
The goal of face style transformation, a developing field at the intersection of image processing, machine learning, and computer vision, is to strengthen and enhance features with no compromising the true nature of an individual. This technology generates precise and realistic modifications that may be performed to a variety of actions by applying complex algorithms, particularly Generative Adversarial Networks (GANs). With the goal to capture the figure of the face, the method involves recognising and recording facial milestones. After that, style transfer methods are used to alter the appearance in accordance with needed characteristics like ageing, communication, or cultural trends. Face border exploration, feature birth, and generative modelling are all vital elements of the fundamental technology behind face style transformation. Facial corner discovery provides an accurate and comprehensive track of the face, providing specific modifications, similar to the use of 468 landmarks in models that include MediaPipe Face Mesh. Point birth involves examining the face structure and associating important traits that define the appearance of the existence today. This data is also used in generative modelling purposes, notably when utilising GANs, to produce realistic, high-quality images that depict the required adjustments. This combination of methods provides transformations that are both contextually appropriate and visually pleasing, making them suited to a wide range of purposes that include virtual pass-ons to entertainment.

Keywords: Generative Adversial Networks (GANs), Machine Learning (ML), MediaPipe, 468 landmarks, image processing.