Authors | Sonal C. Bhangale1, Laxmikant S. Dhamande2, Bhagyashree Ashok Tingare3 , Tarun Dhar Diwan4 , R.A. Kapgate1, P. William5 , Prasad M. Patare2 |
Affiliations |
1Department of Mechatronics Engineering, Sanjivani College of Engineering, Kopargaon, MH, India 2Department of Mechanical Engineering, Sanjivani College of Engineering, Kopargaon, MH, India 3Department of Artificial Intelligence and Data Science, D Y Patil College of Engineering, Akurdi, Pune 4Controller of Examination (COE), Atal Bihari Vajpayee University, Bilaspur, India 5Department of Information Technology, Sanjivani College of Engineering, Kopargaon, MH, India |
Е-mail | bhangalesonalmk@sanjivani.org.in |
Issue | Volume 17, Year 2025, Number 4 |
Dates | Received 03 April 2025; revised manuscript received 18 August 2025; published online 29 August 2025 |
Citation | Sonal C. Bhangale, Laxmikant S. Dhamande, et al., J. Nano- Electron. Phys. 17 No 4, 04028 (2025) |
DOI | https://doi.org/10.21272/jnep.17(4).04028 |
PACS Number(s) | 68.37.Ps, 68.65.Pq |
Keywords | Graphene layer, Atomic structure, Effective Chicken Swarm Guided Recursive NeuroNet (ECS-RNN), Median and wiener filter. |
Annotation |
The qualities of graphene and other atomic materials, coupled with computer vision methods, enable radical improvements in the sensitivity of flaw detection. The data collection process obtained in the methods was high-resolution imaging which applied to capture minute surface details of the atomic materials. The photos are exposed to contemporary feature extraction techniques to highlight and improve the details of the main structural components after undergoing stringent data pretreatment stages like noise reduction and picture standardization. The unique Effective Chicken Swarm Guided Recursive NeuroNet (ECS-RNN) model aims to classify and detect defects by applying smart swarms and deep learning. Trained on the performance metrics achieved in this study, which indicates its capability of producing very high accurate and reliable predictions of 97.3 % F1-score, 98.5 % accuracy and 96.8 % precision. These results indicate advancement in defect detection using the proposed technique and show applicability of machine learning practices in solving very complex problems. The ECS-RNN model reveals substantial improvements in neural network computations, demonstrates its ability to retrieve relevant architectures with minimal erosion, which is advantageous in scenarios where the retrieval of crucial information is of utmost importance. |
List of References |