An intelligent system for content generation
Generation graphical content from a single template image file in art manner is an important but difficult task for artificial neural networks, mostly due to the huge difference between classification existing data and producing new data. Nevertheless artificial intelligent systems capable of genera...
Đã lưu trong:
| Những tác giả chính: | , , , , , , , |
|---|---|
| Định dạng: | Статья |
| Ngôn ngữ: | English |
| Được phát hành: |
CEUR-WS
2018
|
| Những chủ đề: | |
| Truy cập trực tuyến: | https://www.scopus.com/record/display.uri?eid=2-s2.0-85019885850&origin=resultslist&sort=plf-f&src=s&st1=An+intelligent+system+for+content+generation&st2=&sid=29bb6676f1cc3369baba28b866977483&sot=b&sdt=b&sl=59&s=TITLE-ABS-KEY%28An+intelligent+system+for+content+generation%29&relpos=76&citeCnt=1&searchTerm= https://dspace.ncfu.ru/handle/20.500.12258/3022 |
| Các nhãn: |
Thêm thẻ
Không có thẻ, Là người đầu tiên thẻ bản ghi này!
|
| Tóm tắt: | Generation graphical content from a single template image file in art manner is an important but difficult task for artificial neural networks, mostly due to the huge difference between classification existing data and producing new data. Nevertheless artificial intelligent systems capable of generating new content are important scientific task, because their working principles are close to the human thinking processes. Here we introduce an artificial system based on a Deep Neural Network that creates images, audio or 3D-content. In this work, we propose an content generative system for producing content by using pre-trained Deep Neural Network. This is made possible mainly two technical innovations. First, we propose to use different pre-trained neural networks, so that generative system can use optimized network parameters to produce new images. Content generative system and core Deep Neural Network are weakly bound components and we can obtain different system output by core replacement. Second, proposed artificial system can be used not only for image generation, but also for producing audio content and generation 3D-models with target style. |
|---|