Self/Object explores my personal connection with objects through making the experimental portraits in different artistic styles. The objects that I choose include playing card, sand glass, flower, mirror, and spinning top. Those self-portraits reflect my feelings and imagination inspired by those objects. The process of making these self-portraits in different artistic styles also drove me to make a style transfer experiment based on my own artworks. In fact, a great number of computational artists have explored this area who are working on using the model of famous paintings to make the style transfer effects or generate new artworks, but how about using the artistic style of my own artworks to implement the style transfer?
I selected three of my works to experiment with style transfer, here is the video recording of my style transfer website, the website uses the WebCam to capture video and live transfer it into my artistic style:
Here are my original self-portraits and the image of style transfer results:
The self-portraits were all shot by myself and edited in the Adobe Photoshop. For the web-based fast style transfer, I firstly trained the model of my artworks used deeplearn.js. After finishing the machine learning, I imported the models and built the website with ml5.js, p5.js, html, css.