Zero-Shot Learning and its Applications from Autonomous Vehicles to COVID-19 Diagnosis: A Review

Overview

#COVID-19 is not only a global challenge and concern for the medical community but also for Compute Scientists and Computer Vision Researchers. This paper provides with a review on the emerging topic of Zero-Shot Learning which is applicable to both Engineering and Medical Imaging, particularly when we do not have enough datasets such as limited annotated and labelled Chest CT scans of the COVID-19 patients.

Summary

To the best of our knowledge, we are the first who provide with this type of review paper on the concept of zero-shot learning classification which is applicable in medical science, engineering, and other real-world applications, where we have no prior knowledge or enough knowledge for classifying a new concept (e.g. classifying positive cases of COVID-19 from negative ones based on the Chest X-ray)

Author Comments

Dr Mahdi Rezaei, PhD
Dr Mahdi Rezaei, PhD
The University of Leeds
Assistant Professor
PhD in Computer Science
Leeds | United Kingdom
The recent COVID-19 outbreak and pandemic are leading to serious global issues and costs, not only for health organisations but also for almost all type of jobs. Therefore we Computer Science and AI researchers, particularly in the field of Computer vision and Medical Imaging can contribute towards early detection of the positive COVID-19 patients based on the CT-scan images. This paper reviews the emerging concept of Zero-shot learning and object classification techniques, for detecting new objects or concepts of any types e.g. automatic detection for a new concept car which has not been seen before; or a new type of virus (again never seen before), a new type of chest X-ray.Dr Mahdi Rezaei, PhD

Resources

arXiv
https://arxiv.org/abs/2004.14143
Researchgate
https://www.researchgate.net/publication/341035421_Zero-Shot_Learning_and_its_Applications_from_Autonomous_Vehicles_to_COVID-19_Diagnosis_A_Review
Twitter
https://twitter.com/DrMahdiRezaei/status/1256587757197066243

Similar Publications