Implementation of Multi-view learning to a Dynamic Convolutional Neural Network for Object Detection

In the context of developing an academic activity as part of the intern’s postgraduate studies and building a new international research collaboration, the University of Alberta and the Tecnológico de Monterrey propose this research stay project, promoting a strategic initiative of artificial intelligence (AI) and the establishment of Industry 5.0 in fields such as agriculture and biomedicine. The project involves implementing and validating Multi-view learning in a Dynamic Convolutional Neural Network (MVDCNN) for object detection, utilizing different feature fusion strategies to enhance the performance of single-view detector models in terms of mean average precision (mAP). This collaboration aims to generate open-access knowledge that might encourage the development of technological solutions to face current national challenges, such as food security and health care access, addressed by government programs such as Food Sovereignty, Health and Data Science (México), Science and innovation in agriculture, and Health Science and Research (Canada). The awaited outcome of the research resides in the design, implementation, and validation of a multi-view detection model that improves detection rates by applying feature fusion techniques, contributing to further projects, including the intern’s thesis project, and delivering a manuscript for submission to a prestigious conference or journal.

Faculty Supervisor:

Li Cheng

Student:

Partner:

Instituto Tecnológico y de Estudios Superiores de Monterrey (ITESM)

Discipline:

Computer science

Sector:

Artificial Intelligence; Agriculture and Food; Health and Related Sciences and Technology

University:

University of Alberta

Program: