Generative Adversarial Networks with Conditional Neural Movement Primitives for An Interactive Generative Drawing Tool

TL;DR

This paper introduces GAN-CNMP, a novel framework combining GANs with Conditional Neural Movement Primitives, to generate smooth, consistent sketches with limited labeled data, improving shape quality in interactive drawing tools.

Contribution

The paper presents a new GAN-based framework with a novel adversarial loss on CNMPs, enhancing sketch smoothness and shape consistency, trained efficiently with few unlabeled samples.

Findings

Model produces smoother, more consistent sketches.

Effective with limited unlabeled data.

Outperforms baseline models in shape quality.

Abstract

Sketches are abstract representations of visual perception and visuospatial construction. In this work, we proposed a new framework, Generative Adversarial Networks with Conditional Neural Movement Primitives (GAN-CNMP), that incorporates a novel adversarial loss on CNMP to increase sketch smoothness and consistency. Through the experiments, we show that our model can be trained with few unlabeled samples, can construct distributions automatically in the latent space, and produces better results than the base model in terms of shape consistency and smoothness.