Real-Time Style Transfer With Strength Control
This work addresses a practical need for flexible style adjustment in applications like photo editing and media production, but it is incremental as it builds directly on prior real-time style transfer methods.
The paper tackles the problem of enabling smooth, continuous control over stylization strength in real-time style transfer, which was previously limited to fixed styles or required multiple networks. The proposed method achieves this with a single transformer network, improving detail reproduction and reducing artifacts compared to existing alternatives.
Style transfer is a problem of rendering a content image in the style of another style image. A natural and common practical task in applications of style transfer is to adjust the strength of stylization. Algorithm of Gatys et al. (2016) provides this ability by changing the weighting factors of content and style losses but is computationally inefficient. Real-time style transfer introduced by Johnson et al. (2016) enables fast stylization of any image by passing it through a pre-trained transformer network. Although fast, this architecture is not able to continuously adjust style strength. We propose an extension to real-time style transfer that allows direct control of style strength at inference, still requiring only a single transformer network. We conduct qualitative and quantitative experiments that demonstrate that the proposed method is capable of smooth stylization strength control and removes certain stylization artifacts appearing in the original real-time style transfer method. Comparisons with alternative real-time style transfer algorithms, capable of adjusting stylization strength, show that our method reproduces style with more details.