Notice of Violation of IEEE Publication Principles: HopNet based Associative Memory as FC layer in CNN for Odia Character Classification

Notice of Violation of IEEE Publication Principles

“HopNet based Associative Memory as FC Layer in CNN for Odia Character Classification”
by Ramesh Chandra Sahoo, Sateesh Kumar Pradhan, Poonam Tanwar
in the Proceedings of the 10th International Conference on Cloud Computing, Data Science & Engineering (Confluence), January 2020

After careful and considered review of the content and authorship of this paper by a duly constituted expert committee, this paper has been found to be in violation of IEEE’s Publication Principles.

This paper contains copied content from the paper cited below. The original content was copied without attribution (including appropriate references to the original author(s) and/or paper title) and without permission.

“Unsupervised Learning using Pretrained CNN and Associative Memory Bank”
by Qun Liu and Supratik Mukhopadhyay
in the Proceedings of the International Joint Conference on Neural Networks (IJCNN), July 2018


A deep neural network such as convolutional neural network is a popular and most commonly applied technique in image processing for classification for the last few years. The overhead of the feature extraction step will be avoided due to the implicit feature extraction nature of convolutional neural network (CNN) and these extracted features contain substantial information that could be sufficient for an image classification problem. Fully connected (FC) layers in CNN take the results of the last convolution and/or pooling layer and then use them to recognize or classifying images into labels. In this paper, we present an associative memory-based model named Hopfield network as a fully connected layer to store patterns for classification in CNN architecture like LeNet-5. The main purpose of using Hopfield network is to avoid backpropagation as it is a fully connected recurrent network as the state-of-art results which we have obtained are comparable with other models. To measure the performance of the new architecture, we used NIT, Rourkela, Odia characters dataset and compared it with other models for classification.