Four typical network structures for pose-guided person image generation. (a) Top-down method directly concatenates í µí°¼ í µí± , í µí±ƒ í µí± and í µí±ƒ í µí±¡ as input to synthesize í µí°¼ í µí±¡ . (b) Bottom-up latent feature representation method extracts the source appearance feature í µí±“ í µí°ánd target pose feature í µí±“ í µí±ƒ as internal representations. (c) Bottom-up deformation method tries to deform the feature maps of í µí°¼ í µí± and í µí±ƒ í µí± into the target pose í µí±ƒ í µí±¡ . (d) Hybrid method fuses the top-down and bottom-up deformation methods into a framework.

Four typical network structures for pose-guided person image generation. (a) Top-down method directly concatenates í µí°¼ í µí± , í µí±ƒ í µí± and í µí±ƒ í µí±¡ as input to synthesize í µí°¼ í µí±¡ . (b) Bottom-up latent feature representation method extracts the source appearance feature í µí±“ í µí°ánd target pose feature í µí±“ í µí±ƒ as internal representations. (c) Bottom-up deformation method tries to deform the feature maps of í µí°¼ í µí± and í µí±ƒ í µí± into the target pose í µí±ƒ í µí±¡ . (d) Hybrid method fuses the top-down and bottom-up deformation methods into a framework.

Source publication
Preprint
Full-text available
Deep person generation has attracted extensive research attention due to its wide applications in virtual agents, video conferencing, online shopping and art/movie production. With the advancement of deep learning, visual appearances (face, pose, cloth) of a person image can be easily generated or manipulated on demand. In this survey, we first sum...

Similar publications

Article
Full-text available
This article explores key determinants of the intention to work from home (WFH) among U.S. adults in the early phase of the pandemic. Leveraging nationally representative survey data collected in the initial stages of the pandemic, it explores the role of modalities of communication alongside the more frequently studied behavioral, occupational, an...
Article
Full-text available
This study investigates the distance learning experiences of rural teachers during Covid-19 pandemic. The study employed phenomenological qualitative research design. In the study, 8 classroom teachers working in rural areas of Malatya province in eastern Turkey participated in the 2020-2021 academic year. Research data was collected through semi-s...
Article
Full-text available
Background and Aims Cardiac rehabilitation (CR) is traditionally delivered in-person, however, the COVID-19 pandemic provided impetus for alternate offerings such as telehealth. We investigated uptake, barriers, and enablers in a national survey during the pandemic in Australia. Methods We surveyed CR programs between April-June 2021 using profess...
Preprint
Full-text available
Traditionally, video conferencing is a widely adopted solution for telecommunication, but a lack of immersiveness comes inherently due to the 2D nature of facial representation. The integration of Virtual Reality (VR) in a communication/telepresence system through Head Mounted Displays (HMDs) promises to provide users a much better immersive experi...
Article
Full-text available
Lay abstract: Parents of children with autism are known to experience severe hardships related to raising their children. These hardships are exacerbated in low-resource settings internationally where there is very little resource for children and their families, including professionals who provide evidence-based treatment. Mongolia was chosen as...