Neural Doodle is an open-source project that transforms simple doodles into detailed artworks by leveraging deep neural networks. It utilizes semantic style transfer to apply the artistic style of a reference image to a target sketch, enabling the creation of seamless textures, style transfers, and example-based upscaling. The system operates by matching annotated patches from the style image to the target image, facilitating intricate and context-aware artistic renderings.
Key Features
Semantic style transfer using annotated masks
Generation of seamless textures from photographs
Application of artistic styles to sketches and doodles
Seerah Legacy
.png)
.png)
.png)
.png)
.png)
.png)