Back to Blog
![]() We select graphical models from various domains that typically combine textual and graphical elements. In this study, we investigate the suitability of optical character recognition (OCR) as a basis for such a uniformed approach. Notation independence implies that such a uniform approach can only be based on elements commonly present in models of different domains, i.e., text, boxes, and lines. Therefore, to ensure maintenance of multi-domain systems, we need a uniform approach that would be independent from the peculiarities of the notation. ![]() Indeed, these models are created using different modeling notations and it is not plausible to use a multitude of parsers geared toward each and every modeling notation. Only few tools, however, support management of models from different domains. Although these models belong to different domains, the changes in one model can affect other models causing inconsistencies in the entire system. For example, to develop a mechatronic component one might need to combine expertise about mechanics, electronics, and software. The development of systems following model-driven engineering can include models from different domains. Again, the length of solid diagonal lines is greater than or equal 20. Solid diagonal line detection is divided into two parts: left solid diagonal lines and right solid diagonal lines. When image resolution is low, some lines could be detected as two or more lines. This array is used to check if a line segment is a part of another line already detected before. Lines which are smaller than 20 pixels are saved in another 2- dimensional array called " checked solid lines ". All solid straight lines that are detected are stored in a 2-dimensional array called " lines ". Experimenting with many different values for this lenght showed that using a value of 20 produced the best result. This value is choosen to reduce problems of detecting symbols lines as straight lines. The tool starts detecting solid straight lines whose length is greater than or equal to 20 pixels. Solid straight lines detection contains horizontal lines and vertical lines. ![]() a) Solid line detection: This step is divided into two parts: solid straight lines detection and solid diagonal lines detection. Relationships detection is divided into 3 parts: solid line detection, dashed line detection and connecting segments of lines. We performed a validation which shows that Img2UML successfully handles a large class of UML Class images.Īnother presentation of the generalization relationship in class diagram In this process we just detect which classes have relationships with each other without detection of the kind of the relationships. The Img2UML tool exports UML Class models into XMI files which can be read with the StarUML CASE tool. The Img2UML extracts UML Class models from images such that these models can be loaded into CASE tools for further study. In this paper we propose the Img2UML tool to solve this problem. Current CASE tools cannot recognize information from images. These images do not include the model information that is available for UML models when saved in a CASE (Computer Automated Software Engineering) tool format or the XML-based version of the UML interchange format: XMI. A big problem for studies of UML models is that UML models are published mostly in image formats (such as JPEG). Given their limited availability, we resort to collecting UML models from internet. Ideally, researchers want to study UML models from documentations from industrial software projects. Studying UML models is important to understand their effectiveness in software development. In this field, UML is considered the standard for software modeling. Software modeling is an active field of research.
0 Comments
Read More
Leave a Reply. |