Dance energy style transfer using optical flow pattern and image-to-image translation networks

Показати скорочений опис матеріалу

dc.contributor.author Petryshyn, Sofiia
dc.date.accessioned 2024-02-14T09:04:28Z
dc.date.available 2024-02-14T09:04:28Z
dc.date.issued 2022
dc.identifier.citation Petryshyn, Sofiia. Dance energy style transfer using optical flow pattern and image-to-image translation networks / Petryshyn, Sofiia; Supervisor: Roman Vey, Lyubomyr Senyk; Ukrainian Catholic University, Department of Computer Sciences. – Lviv: 2022. – 47 p. uk
dc.identifier.uri https://er.ucu.edu.ua/handle/1/4405
dc.language.iso en uk
dc.title Dance energy style transfer using optical flow pattern and image-to-image translation networks uk
dc.type Preprint uk
dc.status Публікується вперше uk
dc.description.abstracten Generative models was a topic of interest in a last year’s research. ‘Can Machines be More Creative than Humans?’ - the answer to this question is generative art. General thought that ’Generative art incorporates a self-governed or autonomous system’ is no longer relevant since deep learning techniques have made rapid progress in conditional image generation. This work addresses image-to-image translation problem. Many image-to-image translation problems are ambiguous, as a single input image may correspond to multiple possible outputs. In this work, we aim to model a distribution of possible outputs in a conditional generative modeling setting. We train conditional GAN with additional changes, such as one more channel as an input, different pairs in datasets, and changes in input images sizes. As a result, we propose a framework that efficiently generates energy-flow video visuals from a single input-video where a person is dancing. uk


Долучені файли

Даний матеріал зустрічається у наступних зібраннях

Показати скорочений опис матеріалу

Пошук


Перегляд

Мій обліковий запис