No full-text available
To read the full-text of this research,
you can request a copy directly from the authors.
Sketch Based Database Querying System
Abstract
Humans are more familiar to hand drawn sketches than to any other mode of expressing ideas. Specially, laymen find it easy to understand content represented in visual form than in textual form. For example, UML diagrams are used as a standard technique to visually present designs of software systems; which could be easily understood by novices to software development too. However, there are no standard technique to present database queries using diagrams. There are some notations to represent query execution plan which is used to describe the actual internal process at run-time. Thus, through this research study a novel sketch based query language for database querying is introduced. Initially, a user study was conducted to determine the most appropriate shapes and symbols to be used in the proposed sketch based query language. Subsequently, image processing techniques have been applied to recognize circular shapes, arrows, joined circles and text regions in sketches. Thus, a hand drawn sketch will be automatically translated to SQL queries and relevant data will be retrieved from a database. Furthermore, this research study introduces a novel method to recognize the joined circular shapes.
http://ieeexplore.ieee.org/document/8300391/
We present our recent model of a diagram recognition engine. It extends our previous work which approaches the structural recognition as an optimization problem of choosing the best subset of symbol candidates. The main improvement is the integration of our own text separator into the pipeline to deal with text blocks occurring in diagrams. Second improvement is splitting the symbol candidates detection into two stages: uniform symbols detection and arrows detection. Text recognition is left for postprocessing when the diagram structure is already known. Training and testing of the engine was done on a freely available benchmark database of flowcharts. We correctly segmented and recognized 93.0% of the symbols having 55.1% of the diagrams recognized without any error. Considering correct stroke labeling, we achieved the precision of 95.7 %. This result is superior to the state-of-the-art method with the precision of 92.4 %. Additionally, we demonstrate the generality of the proposed method by adapting the system to finite automata domain and evaluating it on own database of such diagrams.
Diagrams are widely used by software engineers to capture the structure and organization of software systems. The Unified Modeling Language (UML) is a commonly-used notation for such diagrams. We have designed and implemented a system for the on-line recognition of hand drawn UML diagrams. Input comes from an electronic whiteboard, a mouse, or a data tablet. A sophisticated segmentation algorithm groups pen strokes into symbols, coping with drawing inaccuracies that are common in hand drawn input. The system is organized around a retargetable kernel which provides a general front end for on-line recognition of any iconic notation. The kernel is extended with UML specific enhancements to segmentation, as well as UML specific glyph recognizers. A simple and intuitive graphical user interface allows the user to correct segmentation and recognition errors. Relatively messy freehand UML drawings are interpreted properly.
In this paper, we propose that deep neural networks that are thinner than typical image recognition networks can perform sketch recognition effectively. As in other computer vision problems, convolutional neural networks (CNNs) outperform other feature extraction methods significantly in sketch recognition. To date, two CNN structures have been proposed for sketch recognition, as described in [4] and [5], achieving classification accuracies of 74.9% and 75.42%, respectively. Although an easy way to improve performance is to design a deeper network, researchers currently tend to avoid an increase in the number of parameters. We attempted to find a more efficient network structure in terms of trade-off between memory usage and performance. Our experiments suggest that making convolutional layers thinner does not result in a significant drop in performance, which correspond to our expectation that low level features are simpler in sketches than those extracted from images.
Information has become a very important resource of any organization and a great demand exists for speedily inputting the colossal amount of printed and handwritten information present on documents into the computer, which requires a lot of manual labour. Over the past few decades, electronic document management systems have turned out to be 'beneficial to' and 'popular in' the society. Profound research and development has been done in this field. Automatic pattern analysis & recognition systems have been & are being developed to analyze the contents of the documents. An important stage in this procedure is the automatic recognition of diagrams from the document. Diagram recognition being one of the main areas in the field of document analysis & pattern recognition is still an open area for researchers. Diagram recognition is a trivial task for humans, but the development of its computer based solution is tremendously difficult. Various subject-specific methods have been developed. In this paper, we present a novel cognitive approach for the recognition of offline typeset/machine-drawn diagrams by the identification of its symbols. The focus here is to use minimal amount of data for the purpose of recognition and to provide a general system for recognition of various types of diagrams. This involves the development of a Recognition Domain Knowledge Kernel. The algorithm proposed is independent of the size of the symbol used in the diagram. Based on the type of symbols present in the diagram, a diagram classification is made.
Structured diagrams are very prevalent in many document types. Most people who need to create such diagrams use structured graphics editors such as Microsoft Visio. Structured graphics editors are extremely powerful and expressive but they can be cumbersome to use. We have shown through extensive timing experiments that structured diagrams drawn by hand will take only about 10% of the time it takes to draw one using a tool like Visio. This indicates the value of automated recognition of hand-written diagrams. Recently, applications have been developed that use online systems running on pen-input PCs that allow users to create structured diagrams by drawing the diagram on the PC tablet. The progress of offline diagram recognition is still minimal. The objective of this paper is to propose a context-independent off-line diagram recognition system. Our approach utilizes support vector machines for recognition and line primitive extraction by interpretation of line continuation for segmentation.
An Online Hand-Drawn Electric Circuit Diagram Recognition System Using Hidden Markov Models
- Y Zhang
- C Viard-Gaudin
- L Wu
Y. Zhang, C. Viard-Gaudin and L. Wu, "An Online Hand-Drawn
Electric Circuit Diagram Recognition System Using Hidden Markov
Models," in International Symposium on Information Science and
Engineering, Shanghai, 2008.
On-line sketch recognition for course of action diagrams
- X Wang
- H Xu
- H Wang
X. Wang, H. Xu and H. Wang, "On-line sketch recognition for course of
action diagrams," in IEEE International Conference on Mechatronics
and Automation, Xi'an, 2010.