AE-CNN Based Supervised Image Classification

Chandra, G. and Challa, M. (2021) AE-CNN Based Supervised Image Classification. In: 5th International Conference on Computer Vision and Image Processing, CVIP 2020, 4 December 2020 to 6 December 2020, Prayagraj.

Full text not available from this repository. (Request a copy)


Point of Care Ultrasound (PoCUS) imaging is an important tool in detecting lung consolidations and tissue sliding, and hence has a potential to identify the onset of novel-CoVID-19 attack in a person. Of late, Convolutional Neural Network (CNN) architectures have gained popularity in improving the accuracy of the predictions. Motivated by this, in this paper, we introduce a CNN based Auto Encoder (AE-CNN) for a better representation of the features to get an accurate prediction. While most of the existing models contain ‘fully connected’ (FC) layers, in our work, we use only convolutional layers instead of FC layers before the output layer, which helps us in achieving a less training time of the model. Moreover, fully connected layers of a network can not learn the patterns in an image as much as convolutional layers can. This is the main advantage of our model over its existing counterparts. We demonstrate that our model detects the lung abnormalities in the ultrasound images with an accuracy of 96.6%.

[error in script]
IITH Creators:
IITH CreatorsORCiD
-NA-, -NA--NA-
Item Type: Conference or Workshop Item (Paper)
Additional Information: Series Title: Communications in Computer and Information Science
Uncontrolled Keywords: Convolution Neural Network, Extreme Gradient Boosting (XGBoost), Image processing, Lung ultrasound, Machine learning
Subjects: Computer science
Divisions: Department of Computer Science & Engineering
Depositing User: Mrs Haseena VKKM
Date Deposited: 09 Nov 2021 09:21
Last Modified: 09 Nov 2021 09:21
Publisher URL:
OA policy:
Related URLs:

Actions (login required)

View Item View Item
Statistics for RAIITH ePrint 8900 Statistics for this ePrint Item