Rasterization and vectorization are dual problems in computer graphics. We achieve this by a component-wise path initialization method and a novel Unsigned Distance guided Focal loss function (UDF loss).īesides, to mitigate the self-interaction issue, which always occurs in the optimization course, we present a novel Self-Crossing loss (Xing loss) by adding constraints to the control points optimization.
The key insight behind this idea is that simply minimizing the vectorization error ( e.g., MSE loss between an input raster image and rendered vector graphics) for optimization would lead to a color mean error.
In each step, we are in pursuit of maximizing the topology exploration rather than only minimizing the pixel-wise difference. Moreover, LIVE enjoys an intuitive and succinct learning course. This property helps us to escape the regime of particular domains like fonts and emojis, and bypass the difficulty of SVG dataset collection or generalization. In this paper, we introduce a Layer-wise Image VEctorization method, termed as LIVE, to translate a raster image to vector graphics ( i.e., SVG) with layer-wise representation.ĭifferent from previous works, LIVE is model-free and requires no shape primitive labels. Hence, a simple yet effective method is desired in the community to capture the layer-wise representation for image-to-vector translation. The other line that considers segmentation pre-processing method requires heavy pre-processing operations and would segment high-contrast texture into multiple small regions, resulting in redundancy. The first line of work learns to explore the geometric information of fonts or emojis but cannot be generalized to broad domains. Some methods attempt to resolve this dilemma by either focusing on particular simple datasets or employing a segmentation pre-processing method, but each one has its own drawbacks and subtleties. The missing of such information always incurs inadequate learning of vectorization and requires superfluous shape primitives to make up. These methods, despite their promising vectorization and generation ability, have always overlooked the topological information hidden behind the raster images. In the last few years, we have witnessed various achievements in image-to-vector translation, mainly due to advances in two technical directions: building powerful generation models, and employing decent differentiable rendering methods. Initiates human editable SVGs for both designers and other downstreamĪpplications. With the help of this newly learned topology, LIVE LIVE presents more plausible vectorized forms than prior works and can be With the layer-wise framework, newly designed loss functions, andĬomponent-wise path initialization technique. We progressively add new bezier paths and optimize these paths With layer-wise structures that are semantically consistent with human Simultaneously maintain its image topology. Image Vectorization, namely LIVE, to convert raster images to SVGs and Layer-wise topology and fundamental semantics in images are still not well The generated SVGs also contain complex and redundant shapes thatĪre not quite convenient for further editing.
However, deep models cannot be easily generalized to out-of-domain Interpolation of vector graphs and demonstrate a better topology of generating Recent advanced deep learning-based models achieve vectorization and semantic Vectorization, the reverse path of rasterization, remains a major challenge. Image rasterization is a mature technique in computer graphics, while image