An efficient graph-based recognizer for hand-drawn symbols

Lee, WeeSan; Kara, Levent Burak; Stahovich, Thomas F.

doi:10.1016/j.cag.2007.04.007

Cited by 41 publications

(23 citation statements)

References 16 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…They used this language to perform automatic symbol recognition. The attributed relational graph (ARG) is an excellent statistical model to describe both geometry and topology of a symbol [2], and is insensitive to orientation, scaling, and drawing order. The advantage of structural methods is distinguishing similar shapes.…”

Section: Related Workmentioning

confidence: 99%

See 1 more Smart Citation

On-Line Sketch Recognition Using Direction Feature

Wu¹,

Yu²,

Lai³

2013

Human-Computer Interaction – INTERACT 2013

View full text Add to dashboard Cite

Abstract. Sketch recognition is widely used in pen-based interaction, especially as the increasing popularity of devices with touch screens. It can enhance human-computer interaction by allowing a natural/free form of interaction. The main challenging problem is the variability in hand drawings. This paper presents an on-line sketch recognition method based on the direction feature. We also present two feature representations to train a classifier. We support our case by experimental results obtained from the NicIcon database. A recognition rate of 97.95% is achieved, and average runtime is 97.6ms using a Support Vector Machine classifier. Keywords: Sketched symbol recognition, NicIcon database, multi-stroke shapes IntroductionSketching is a natural form of human communication. Sketch-based interaction is a fast and efficient means of capturing information by automatically interpreting handdrawn sketches and can be an important part of the early design process, where it can help people explore rough ideas and solutions in an informal environment. Sketch recognition refers to the recognition of predefined symbols or free-form drawings(e.g., an unconstrained circuit drawing); in the latter case, the recognition task is generally preceded by segmentation in order to locate individual symbols. This paper focuses on the recognition of hand-drawn isolated symbols. With the growing popularity of devices with touch screens, there is increasing interest in building sketch-based user interfaces. However, many challenges remain in terms of intra-class compactness and interclass separation due to the variability of sketching. Because it is likely that different people will have different drawing styles, such as the stroke-order, -number, and nonuniform scaling , as well as complex local shifts. Moreover, the style of the same individual at different times may differ. A good recognition algorithm should place few drawing constraints on users. Related research is that of handwriting recognition, such as handwritten digit and Chinese character recognition with many effective algorithms.

show abstract

Section: Related Workmentioning

confidence: 99%

“…Then four vectors can form a final shape descriptor, which is the feature representation of a symbol. Initialize the probability vector v as v( j )=0, 2 1, 2,..., j k = .…”

Section: Feature Representationmentioning

confidence: 99%

On-Line Sketch Recognition Using Direction Feature

Wu¹,

Yu²,

Lai³

2013

Human-Computer Interaction – INTERACT 2013

View full text Add to dashboard Cite

show abstract

“…Also, it is not well suited to recognize "messy" gestures like a scratch-out, commonly used for erasing-like actions. Lee et al [8] present a trainable graph-based recognizer that is insensitive to orientation, scale and drawing direction and is able to recognize multi-stroke gestures. Since the recognizer uses statistical models to define symbols, it handles the small variations associated with hand-drawn gestures very well.…”

Section: A Sketch Recognizersmentioning

confidence: 99%

SketchyDynamics: A Library for the Development of Physics Simulation Applications with Sketch-Based Interfaces

Costa¹,

Pereira²

2013

IJIMAI

View full text Add to dashboard Cite

-Sketch-based interfaces provide a powerful, natural and intuitive way for users to interact with an application. By combining a sketch-based interface with a physically simulated environment, an application offers the means for users to rapidly sketch a set of objects, like if they are doing it on piece of paper, and see how these objects behave in a simulation. In this paper we present SketchyDynamics, a library that intends to facilitate the creation of applications by rapidly providing them a sketch-based interface and physics simulation capabilities. SketchyDynamics was designed to be versatile and customizable but also simple. In fact, a simple application where the user draws objects and they are immediately simulated, colliding with each other and reacting to the specified physical forces, can be created with only 3 lines of code. In order to validate SketchyDynamics design choices, we also present some details of the usability evaluation that was conducted with a proof-of-concept prototype.

show abstract

“…Its objective is the recognition of the graphical primitives (such as lines and arcs) composing the strokes. Stroke segmentation can be used for a variety of objectives, including symbol [16,4] and full diagram [3] recognition.…”

Section: Introductionmentioning

confidence: 99%

RankFrag: A Machine Learning-Based Technique for Finding Corners in Hand-Drawn Digital Curves

Costagliola

Rosa

Fuccella

2015

International Conferences on Distributed Multimedia Systems

View full text Add to dashboard Cite

We describe RankFrag: a technique which uses machine learning to detect corner points in hand-drawn digital curves. RankFrag classifies the stroke points by iteratively extracting them from a list of corner candidates. The points extracted in the last iterations are said to have a higher rank and are more likely to be corners. The technique has been tested on three different datasets described in the literature. We observed that, considering both accuracy and efficiency, RankFrag performs better than other state-of-art techniques.

show abstract

An efficient graph-based recognizer for hand-drawn symbols

Cited by 41 publications

References 16 publications

On-Line Sketch Recognition Using Direction Feature

On-Line Sketch Recognition Using Direction Feature

SketchyDynamics: A Library for the Development of Physics Simulation Applications with Sketch-Based Interfaces

RankFrag: A Machine Learning-Based Technique for Finding Corners in Hand-Drawn Digital Curves

Contact Info

Product

Resources

About