Abstract:Freehand sketching is a dynamic process where points are sequentially sampled and grouped as strokes for sketch acquisition on electronic devices. To recognize a sketched object, most existing methods discard such important temporal ordering and grouping information from human and simply rasterize sketches into binary images for classification. In this paper, we propose a novel single-branch attentive network architecture RNN-Rasterization-CNN (Sketch-R2CNN for short) to fully leverage the dynamics in sketches for recognition. Sketch-R2CNN takes as input only a vector sketch with grouped sequences of points, and uses an RNN for stroke attention estimation in the vector space and a CNN for 2D feature extraction in the pixel space respectively. To bridge the gap between these two spaces in neural networks, we propose a neural line rasterization module to convert the vector sketch along with the attention estimated by RNN into a bitmap image, which is subsequently consumed by CNN. The neural line rasterization module is designed in a differentiable way to yield a unified pipeline for end-to-end learning. We perform experiments on existing large-scale sketch recognition benchmarks and show that by exploiting the sketch dynamics with the attention mechanism, our method is more robust and achieves better performance than the state-of-the-art methods.

A Method of Fast On-Line Graphics Recognition and Regularization

On-Line Graphics Recognition

An on-line sketch recognition algorithm for composite shape

Pen-based gesture design and recognition for sketch editing

An online sketching and gesture editing system for conceptual design

Sketch-based 3D Modeling by Aligning Outlines of an Image

A Structural Method for Online Sketched Symbol Recognition.

Sketch-R2CNN: an Attentive Network for Vector Sketch Recognition

Study on User Adaptation for On-Line Sketch Recognition

On-Line Sketch Recognition Using Direction Feature

On-Line Sketchy Graphic Recognition Based on Adaptive HMM

An Efficient Primitive-Based Method to Recognize Online Sketched Symbols with Autocompletion

User Adaptation For Online Sketchy Shape Recognition

Research on Freehand Sketchy Graphics Recognition

Online Composite Sketchy Shape Recognition Using Dynamic Programming

SS-based Sketch Recognition for Graphics of Traffic Accident

Incremental On-Line Sketchy Shape Recognition with Dynamic Modeling and Relevance Feedback

An Online Multi-Stroke Sketch Recognition Method Integrated with Stroke Segmentation

Online Diagramming Recognition Based on Automatic Stroke Parsing and Bayesian Classifier

DyGAT: Dynamic Stroke Classification of Online Handwritten Documents and Sketches

A fast 2D shape recovery approach by fusing features and appearance.