Software algorithms evaluation platform using a Docker-based framework

Abdel Naby, Acil Ramadan Ibrahim (2021) Software algorithms evaluation platform using a Docker-based framework. Masters thesis, Memorial university of Newfoundland.

[img] [English] PDF - Accepted Version
Available under License - The author retains copyright ownership and moral rights in this thesis. Neither the thesis nor substantial extracts from it may be printed or otherwise reproduced without the author's permission.

Download (6MB)


Automating the evaluation of algorithms is a complex task that involves many dependencies, e.g., the inputs, outputs, the compilers, interpreters, execution environments, libraries, operating systems, and hardware dependencies. Each algorithm requires specific performance evaluation criteria to evaluate it during and after execution. This work presents a ADESA (Automatic Deployment and Evaluation of Software Algorithms) framework based on a Docker framework for automating the execution of the algorithms provided by the user and automating the evaluation of these algorithms. The framework is applied to the automatic evaluation of image processing algorithms (AEIPA). AEIPA automates the processes of searching for a matching data set or uploading new data sets, uploading a new image processing algorithm (IPA) that has been developed using any programming language, searching for matching IPAs, executing all those IPAs with one or more data set, and comparing the results of the executed IPAs using each data set. As a case study, two face detection algorithms that implement the Haar Cascade classifier using C with OpenCV and Python with OpenCV, have been used to evaluate AEIPA. Results from this work confirm that AEIPA supports the execution of different programming languages using Docker images.

Item Type: Thesis (Masters)
Item ID: 15714
Additional Information: Includes bibliographical references (pages 91-104)
Keywords: evaluation, automation, deployment, software, algorithms, image processing, model-view-controller framework, data access object, data transfer object
Department(s): Engineering and Applied Science, Faculty of
Date: May 2021
Date Type: Submission
Digital Object Identifier (DOI):
Library of Congress Subject Heading: Algorithms; Automation; Intelligent agents (Computer software); Image processing; Application software

Actions (login required)

View Item View Item


Downloads per month over the past year

View more statistics