Combined Cloud-Based Inference System for the Classification of COVID-19 in CT-Scan and X-Ray Images.
New Gener Comput
; : 1-24, 2022 Nov 20.
Article
in English
| MEDLINE | ID: covidwho-2294470
ABSTRACT
In the past few years, most of the work has been done around the classification of covid-19 using different images like CT-scan, X-ray, and ultrasound. But none of that is capable enough to deal with each of these image types on a single common platform and can identify the possibility that a person is suffering from COVID or not. Thus, we realized there should be a platform to identify COVID-19 in CT-scan and X-ray images on the fly. So, to fulfill this need, we proposed an AI model to identify CT-scan and X-ray images from each other and then use this inference to classify them of COVID positive or negative. The proposed model uses the inception architecture under the hood and trains on the open-source extended covid-19 dataset. The dataset consists of plenty of images for both image types and is of size 4 GB. We achieved an accuracy of 100%, average macro-Precision of 100%, average macro-Recall of 100%, average macro f1-score of 100%, and AUC score of 99.6%. Furthermore, in this work, cloud-based architecture is proposed to massively scale and load balance as the Number of user requests rises. As a result, it will deliver a service with minimal latency to all users.
Full text:
Available
Collection:
International databases
Database:
MEDLINE
Type of study:
Prognostic study
Language:
English
Journal:
New Gener Comput
Year:
2022
Document Type:
Article
Affiliation country:
S00354-022-00195-x
Similar
MEDLINE
...
LILACS
LIS