The Pose-guided person image synthesis task aims to render a person’s image with a desired pose and appearance. Specifically, the appearance is defined by a given source image and the pose by a set of keypoints. Having control over the synthesized person images in terms of pose and style is an important requisite for applications such as e-commerce, virtual reality, metaverse and content generation for the entertainment industry. Read our paper here.

We introduced an approach, named GMNR, that focuses at efficiently generating 3D-aware high-resolution images that are view-consistent across multiple camera pose. Qualitative and quantitative experiments on three datasets demonstrate the merits of our contributions, leading to favorable performance in terms of image generation quality and computational efficiency, compared to existing works.

Automatic handwritten text generation can be beneficial for people having disabilities or injuries that prevent them from writing, adapting an author's writing style or gathering additional data for training deep learning-based handwritten text recognition models. Here, we investigate the problem of realistic handwritten text generation of text sequences with arbitrary length and diverse calligraphic attributes representing writing styles of a writer. For more visit our project page here.

Humans have an outstanding ability to easily communicate and express abstract ideas and emotions through sketch drawings. Creative sketching or doodling is an expressive activity, where imaginative and previously unseen depictions of everyday visual objects are drawn. Creative sketch image generation is a challenging vision problem, where the task is to generate diverse, yet realistic creative sketches possessing the unseen composition of the visual-world objects. Please visit our project page here.