The authors of a new "sketch simplification" program hope to make neural networks more useful to artists frustrated by the choppy results of existing automated line tools.
Apps such as Adobe Creative Suite provide functions to turn pencil drawings into vectors, but the sketches have to be tight and the resulting "inks" often need a lot of cleanup. Though other researchers and developers have applied neural network to the job, Edgar Simo-Serra writes that their model gets more meaningful and human results.
Our model is based on a fully convolutional neural network. We input the model a rough sketch image and obtain as an output a clean simplified sketch. This is done by processing the image with convolutional layers, which can be seen as banks of filters that are run on the input. While the input is a grayscale image, our model internally uses a much larger representation. We build the model upon three different types of convolutions: down-convolution, halves the resolution by using a stride of two; flat-convolutional, processes the image without changing the resolution; and up-convolution, doubles the resolution by using a stride of one half. This allows our model to initially compress the image into a smaller representation, process the small image, and finally expand it into the simplified clean output image that can easily be vectorized.
The source code hasn't been made available, but Simo-Serra's other works and collaborations are equally fascinating and already downloadable.
