3 results
Search Results
Now showing 1 - 3 of 3
Item Depth-aware Neural Style Transfer(Association for Computing Machinery, Inc (ACM), 2017) Liu, Xiao-Chang; Cheng, Ming-Ming; Lai, Yu-Kun; Rosin, Paul L.; Holger Winnemoeller and Lyn BartramNeural style transfer has recently received signi cant a ention and demonstrated amazing results. An e cient solution proposed by Johnson et al. trains feed-forward convolutional neural networks by de ning and optimizing perceptual loss functions. Such methods are typically based on high-level features extracted from pre-trained neural networks, where the loss functions contain two components: style loss and content loss. However, such pre-trained networks are originally designed for object recognition, and hence the high-level features o en focus on the primary target and neglect other details. As a result, when input images contain multiple objects potentially at di erent depths, the resulting images are o en unsatisfactory because image layout is destroyed and the boundary between the foreground and background as well as di erent objects becomes obscured. We observe that the depth map e ectively re ects the spatial distribution in an image and preserving the depth map of the content image a er stylization helps produce an image that preserves its semantic content. In this paper, we introduce a novel approach for neural style transfer that integrates depth preservation as additional loss, preserving overall image layout while performing style transfer.Item Mixed Illumination Analysis in Single Image for Interactive Color Grading(Association for Computing Machinery, Inc (ACM), 2017) DuchĂȘne, Sylvain; Aliaga, Carlos; Pouli, Tania; PĂ©rez, Patrick; Holger Winnemoeller and Lyn BartramColorists often use keying or rotoscoping tools to access and edit particular colors or parts of the scene. Although necessary, this is a time-consuming and potentially imprecise process, as it is not possible to fully separate the influence of light sources in the scene from the colors of objects and actors within it. To simplify this process, we present a new solution for automatically estimating the color and influence of multiple illuminants, based on image variation analysis. Using this information, we present a new color grading tool for simply and interactively editing the colors of de- tected illuminants, which fits naturally in color grading workflows. We demonstrate the use of our solution in several scenes, evaluating the quality of our results by means of a psychophysical study.Item Pigment-Based Recoloring of Watercolor Paintings(Association for Computing Machinery, Inc (ACM), 2017) Aharoni-Mack, Elad; Shambik, Yakov; Lischinski, Dani; Holger Winnemoeller and Lyn BartramThe color palette used by an artist when creating a painting is an important tool for expressing emotion, directing attention, and more. However, choosing a palette is an intricate task that requires considerable skill and experience. In this work, we introduce a new tool designed to allow artists to experiment with alternative color palettes for existing watercolor paintings. This could be useful for generating alternative renditions for an existing painting, or for aiding in the selection of a palette for a new painting, related to an existing one. Our tool first estimates the original pigment-based color palette used to create the painting, and then decomposes the painting into a collection of pigment channels, each corresponding to a single palette color. In both of these tasks, we employ a version of the Kubelka-Munk model, which predicts the reflectance of a given mixture of pigments. Each channel in the decomposition is a piecewise-smooth map that specifies the concentration of one of the colors in the palette across the image. Another estimated map specifies the total thickness of the pigments across the image. The mixture of these pigment channels, also according to the Kubelka- Munk model, reconstructs the original painting. The artist is then able to manipulate the individual palette colors, obtaining results by remixing the pigment channels at interactive rates.