Performance of Image Recognition with Machine Vision: A Systematic Review (#839)

Authors

Lopez-Carreño, Joseph

Calvo-Lavado, Cristhian

Zarate-Perez, Eliseo

Abstract

The objective of this work was to identify the existing techniques, applications, equipment, and technologies applied in the recognition of images with artificial vision through a systematic review of the literature during the period 2020 – 2022. For the review work, the statement was used PRISMA in the selection and analysis of the 142 articles obtained from the scientific databases EBSCO, Engineering Source, ProQuest and ScienceDirect. Studies that were not directly related to the proposed objectives were discarded, identifying 28 articles for full-text review. The review results strongly suggest that Hopfield-type convolutional artificial neural networks are highly effective in performing image recognition and classification tasks. In the same way, the combination of technological tools such as YOLO, Roboflow, Python and OpenCV show that image processing and deep learning are driving new applications that improve the various performance metrics in these tasks. Therefore, artificial vision, unlike technologies that incorporate electronic devices with sensors, allows an interpretation of the environment with a higher degree of representation of reality, being robust to the complexity of data processing.