the main objective of this work, is the comparative study, design, and development of techniques towards a more efficient identification of existence of objects, the recognition of spatial localization in cases of mul...
详细信息
ISBN:
(纸本)9781665427425
the main objective of this work, is the comparative study, design, and development of techniques towards a more efficient identification of existence of objects, the recognition of spatial localization in cases of multiple identification, as well as the solution of placement problems of objects from one fixed location to another, withthe same object characteristics. Two state of the art approaches are considered to achieve these goals. On one hand the TensorFlow framework and on the other hand the OpenCV library. Aiming towards a comprehensive study, we will create a custom demo that will include parameters such as three types of geometric shapes, specific peripheral colors for the three types of shapes, random placement of the shapes on the Cartesian plane and a camera for video capturing at a constant angle. thus, guided by the common working pattern, a thorough and objective comparison of the performances of the two methodologies in the field of real-time object identification is achieved. this objective is of great technological interest, as mechanical vision has managed to be assimilated by more and more scientific fields, rendering as a prerequisite in the common benefit of their leapfrog development, but also with many possibilities of particularly important extensions. During the development of this paper, both an algorithmic approach has been applied through the OpenCV library and the object-oriented Python programming language, as well as the creation of our own object detection classifier model through the TensorFlow framework. In this way the developed components make the most of the two tools and offering real-time results with exceptional accuracy, through the custom demo in a common and objective framework. Finally, an in-depth analysis of the performance of the two approaches is offered, listing boththeir positive features and their black spots in multiple scenarios beyond the main demo, in order to cover a wide range of factors which can possibly
Development cycle of information systems, especially maintenance and modifications, is still time-consuming and expensive. this paper presents a new programming paradigm, the Configuration-orientedprogramming (COP), ...
详细信息
Development cycle of information systems, especially maintenance and modifications, is still time-consuming and expensive. this paper presents a new programming paradigm, the Configuration-orientedprogramming (COP), which introduces a new dimension - configuration - into the object-orientedprogramming paradigm (OOP). this new dimension introduces a new point of view at a configuration through a configuration form that enables the use of inheritance, abstraction, polymorphism, and recursive definition on the configuration. Based on several years of experience in using this new paradigm in real live projects we observed that this new dimension enables a more effective development of complex information systems by associating modules (similar to objects in the OOP) into complex structures (by defining associations among the modules using configuration).
In this paper we investigate a special type of coupling in object-oriented systems. When a method of a class C invokes a method of a class D, the method of C becomes dependent on the representational details of D: the...
详细信息
One of the most pioneering advantages of the Energy Dispersive X-ray Laue Diffraction (EDLD) is the one-shot experiment for investigation of polycrystalline materials. Using a 2D energy-dispersive detector, the EDLD i...
详细信息
暂无评论