End-To-End Computer Vision Framework: An Open-Source Platform for Research and Education

Orhei, Ciprian; Vert, Silviu; Mocofan, Muguras; Vasiu, Radu

doi:10.3390/s21113691

Cited by 16 publications

(5 citation statements)

References 57 publications

(66 reference statements)

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…For our simulation to be reproducible and easy to use we have used the End-to-End Computer Vision Framework -EECVF- [73,74]. EECVF is an adaptable and dynamic framework designed for researching and testing CV concepts, which does not require the user to handle the interconnections throughout the system.…”

Section: Resultsmentioning

confidence: 99%

Dilated Filters for Edge-Detection Algorithms

et al. 2021

Self Cite

View full text Add to dashboard Cite

Edges are a basic and fundamental feature in image processing that is used directly or indirectly in huge number of applications. Inspired by the expansion of image resolution and processing power, dilated-convolution techniques appeared. Dilated convolutions have impressive results in machine learning, so naturally we discuss the idea of dilating the standard filters from several edge-detection algorithms. In this work, we investigated the research hypothesis that use dilated filters, rather than the extended or classical ones, and obtained better edge map results. To demonstrate this hypothesis, we compared the results of the edge-detection algorithms using the proposed dilation filters with original filters or custom variants. Experimental results confirm our statement that the dilation of filters have a positive impact for edge-detection algorithms from simple to rather complex algorithms.

show abstract

Section: Resultsmentioning

confidence: 99%

Dilated Filters for Edge-Detection Algorithms

et al. 2021

Self Cite

View full text Add to dashboard Cite

show abstract

“…), there were buildings which were not recognized, frustrating users. The authors are working on improving the building recognition algorithm for the SH AR mobile app [29,30].…”

Section: Discussion On the Overall Multi-platform User Evaluation 61 General Discussionmentioning

confidence: 99%

User Evaluation of a Multi-Platform Digital Storytelling Concept for Cultural Heritage

et al. 2021

Self Cite

View full text Add to dashboard Cite

Digital storytelling platforms have proven to be a great way of bringing cultural heritage closer to people. What lacks is a deeper understanding of the user experience of such systems, especially in multi-platform digital storytelling. For the last three years, we have been developing a project called Spotlight Heritage Timisoara, which is at its core a digital storytelling platform for the city of Timisoara (Romania), soon to be European Capital of Culture in 2023. The project consists of a website, mobile applications, and interactive museographic and street exhibitions. This paper presents a multi-platform usability evaluation study which employed semi-structured interviews, observations, think-aloud protocol, SUS questionnaire, Net Promoter Score and Product Reaction Cards to gather insights from 105 participants and reveal usability problems in the Spotlight Heritage context. We found out that the four platforms, i.e., interactive touchscreen table, desktop/laptop, mobile and Augmented Reality, have very good usability scores, are considered accessible and useful, work seamlessly together, and create user satisfaction and loyalty, across demographic groups, having the potential to bring people closer to cultural heritage.

show abstract

“…Other approaches include the Data Science Process Model (DASC-PM) (Schulz et al, 2022) and Engineering Data Driven Applications (EDDA) (Hesenius et al, 2019). Another focus is the development of software frameworks for the practical implementation of ML and AI-CV in particular (Orhei et al, 2021;Stephen Gould, 2012). These software implementations are usually based on generalized processes or frameworks that facilitate ML application and provide necessary tools for AI development with notable examples being scikit-learn and TensorFlow.…”

Section: Ai Development Processesmentioning

confidence: 99%

Towards a process for the creation of synthetic training data for AI-computer vision models utilizing engineering data

Schwoch,

Dammann,

Bartl

et al. 2024

Proc. Des. Soc.

View full text Add to dashboard Cite

Artificial Intelligence-based Computer Vision models (AI-CV models) for object detection can support various applications over the entire lifecycle of machines and plants such as monitoring or maintenance tasks. Despite ongoing research on using engineering data to synthesize training data for AI-CV model development, there is a lack of process guidelines for the creation of such data. This paper proposes a synthetic training data creation process tailored to the particularities of an engineering context addressing challenges such as the domain gap and methods like domain randomization.

show abstract

End-To-End Computer Vision Framework: An Open-Source Platform for Research and Education

Cited by 16 publications

References 57 publications

Dilated Filters for Edge-Detection Algorithms

Dilated Filters for Edge-Detection Algorithms

User Evaluation of a Multi-Platform Digital Storytelling Concept for Cultural Heritage

Towards a process for the creation of synthetic training data for AI-computer vision models utilizing engineering data

Contact Info

Product

Resources

About