In a remarkable academic achievement, Dr Banee Bandana Das, Assistant Professor in Department of Computer Science and Engineering and Dr. Saswat Kumar Ram, Assistant Professor in Department of Electronics and Communication Engineering, have made significant contributions to the field of biometric security. Their paper, titled “Person Identification using Autoencoder-CNN Approach with Multitask-based EEG Biometric,” has been published in the esteemed ‘Multimedia Tools and Applications journal, which is recognised as a Q1 journal with an impressive impact factor of 3.6.

This pioneering work showcases a novel approach to person identification using electroencephalogram (EEG) data. The research leverages the power of Autoencoder-CNN models combined with multitask learning techniques to enhance the accuracy and reliability of EEG-based biometric systems.

The publication of this paper not only underscores the high-quality research conducted at SRM University-AP but also places the institution at the forefront of innovative developments in biometric technology. It is a testament to the university’s commitment to advancing scientific knowledge and providing its faculty with a platform to impact the global research community positively.

Abstract

In this research paper, we propose an unsupervised framework for feature learning based on an autoencoder to learn sparse feature representations for EEG-based person identification. Autoencoder and CNN do the person identification task for signal reconstruction and recognition. Electroencephalography (EEG) based biometric system is vesting humans to recognise, identify and communicate with the outer world using brain signals for interactions. EEG-based biometrics are putting forward solutions because of their high-safety capabilities and handy transportable instruments. Motor imagery EEG (MI-EEG) is a maximum broadly centered EEG signal that exhibits a subject’s motion intentions without real actions. The Proposed framework proved to be a practical approach to managing the massive volume of EEG data and identifying the person based on their different task with resting states.

The title of Research Paper in the Citation Format

Person identification using autoencoder-CNN approach with multitask-based EEG biometric. Multimedia Tools Appl (2024).

Practical implementation/social implications of the research

1. To develop a personal identification system using MI-EEG data.
2. This work is about an Autoencoder-CNN-based biometric system with EEG motor imagery inputs for dimensionality reduction and denoising (extracting original input from noisy data).
3. The designed Autoencoder-CNN-based biometric architecture to model MI-EEG signals is efficient for cybersecurity applications.

Collaborations

1. IIITDM, Kurnool, India
2. National Institute of Technology, Rourkela, India
3. University of North Texas, Denton, USA

Future Research Plan

In the future, different deep learning and machine learning methods can be merged to explore better performance in this EEG-based security field and other signal processing areas. We will investigate the robustness deep learning architectures to design a multi-session EEG biometric system.

Link to The Article

Continue reading →

The Department of Computer Science and Engineering is pleased to announce an extraordinary research paper titled “A Deep Transfer Learning Model for Green Environment Security Analysis in Smart City“, authored by Dr Sambit Kumar Mishra, Assistant Professor, was published in the Journal of King Saud University – Computer and Information Sciences that falls within the Q1 quartile with an Impact Factor (IF) of 6.9. The study introduces a model to automatically classify high-resolution scene images for environmental conservation in smart cities. By enhancing the training dataset with spatial patterns, the model improves green resource management and personalised services. It also demonstrates the effectiveness of LULC classification in smart city environments using transfer learning. Data augmentation techniques improve model performance, and optimisation methods enhance efficiency, contributing to better environmental management.

Abstract

The research addresses the importance of green environmental security in smart cities and proposes a morphologically augmented fine-tuned DenseNet121 (MAFDN) model for Land Use and Land Cover (LULC) classification. This model aims to automate the categorisation of high spatial resolution scene images to facilitate green resource management and personalised services.

Dr Mishra collaborated with Dr Rasmita Dash and Madhusmita Sahu from SoA Deemed to be University, India, as well as Mamoona Humayun, Majed Alfayad, and Mohammed Assiri from universities in Saudi Arabia.

His plans include optimising the model using pruning methods to create lightweight scene classification models for resolving challenges in LULC datasets.

Link to the article