Aditya Gupta

• Ahuja, Bharti; Doriya, Rajesh; Salunke, Sharad; Hashmi, Mohammad Farukh & Gupta, Aditya (2023). IoT-Based Multi-Dimensional Chaos Mapping System for Secure and Fast Transmission of Visual Data in Smart Cities. IEEE Access. ISSN 2169-3536. 11, p. 104930–104945. doi: 10.1109/ACCESS.2023.3318014. Full text in Research Archive Show summary

  A “smart city” sends data from many sensors to a cloud server for local authorities and the public to connect. Smart city residents communicate mostly through images and videos. Many image security algorithms have been proposed to improve locals’ lives, but a high-class redundancy method with a small space requirement is still needed to acquire and protect this sensitive data. This paper proposes an IoT-based multi-dimensional chaos mapping system for secure and fast transmission of visual data in smart cities, which uses the five dimensional Gauss Sine Logistic system to generate hyper-chaotic sequences to encrypt images. The proposed method also uses pixel position permutation and Singular Value Decomposition with Discrete fractional cosine transform to compress and protect the sensitive image data. To increase security, we use a chaotic system to construct the chaotic sequences and a diffusion matrix. Furthermore, numerical simulation results and theoretical evaluations validate the suggested scheme’s security and efficacy after compression encryption.

• Rani, Gundala Jhansi; Hashmi, Mohammad Farukh & Gupta, Aditya (2023). Surface Electromyography and Artificial Intelligence for Human Activity Recognition - A Systematic Review on Methods, Emerging Trends Applications, Challenges, and Future Implementation. IEEE Access. ISSN 2169-3536. 11, p. 105140–105169. doi: 10.1109/ACCESS.2023.3316509. Full text in Research Archive Show summary

  Human activity recognition (HAR) has become increasingly popular in recent years due to its potential to meet the growing needs of various industries. Electromyography (EMG) is essential in various clinical and biological settings. It is a metric that helps doctors diagnose conditions that affect muscle activation patterns and monitor patients’ progress in rehabilitation, disease diagnosis, motion intention recognition, etc. This review summarizes the various research papers based on HAR with EMG. Over recent years, the integration of Artificial Intelligence (AI) has catalyzed remarkable advancements in the classification of biomedical signals, with a particular focus on EMG data. Firstly, this review meticulously curates a wide array of research papers that have contributed significantly to the evolution of EMG-based activity recognition. By surveying the existing literature, we provide an insightful overview of the key findings and innovations that have propelled this field forward. It explore the various approaches utilized for preprocessing EMG signals, including noise reduction, baseline correction, filtering, and normalization, ensure that the EMG data is suitably prepared for subsequent analysis. In addition, we unravel the multitude of techniques employed to extract meaningful features from raw EMG data, encompassing both time-domain and frequency-domain features. These techniques are fundamental to achieving a comprehensive characterization of muscle activity patterns. Furthermore, we provide an extensive overview of both Machine Learning (ML) and Deep Learning (DL) classification methods, showcasing their respective strengths, limitations, and real-world applications in recognizing diverse human activities from EMG signals. In examining the hardware infrastructure for HAR with EMG, the synergy between hardware and software is underscored as paramount for enabling real-time monitoring. Finally, we also discovered open issues and future research direction that may point to new lines of inquiry for ongoing research toward EMG-based detection.

• Ahuja, Bharti; Doriya, Rajesh; Salunke, Sharad; Hashmi, Mohammad Farukh; Gupta, Aditya & Bokde, Neeraj Dhanraj (2023). HDIEA: high dimensional color image encryption architecture using five-dimensional Gauss-logistic and Lorenz system. Connection Science. ISSN 0954-0091. 35(1). doi: 10.1080/09540091.2023.2175792. Full text in Research Archive Show summary

  The work presented here is a high dimensional color image encryption architecture (HDIEA) founded on the Lorenz-Gauss-Logistic (LGL) encryption algorithm. The primary objective is to demonstrate that both the proposed novel five-dimensional (5D) Gauss-Logistic and four-dimensional (4D) Lorenz system are operating in a hyper-chaotic condition. The visual study of their most important characteristics, such as the sensitivity of the starting value of both maps and the Lyapunov exponent of the 5D Gauss Logistic map, is carried out. The Runge–Kutta technique is used to discretise the Lorenz system in order to construct a pseudo-random sequence generation for the control parameter that has a greater degree of randomness. The 5D Gauss-Logistic system is then selected to serve as the principal hyper-chaotic mapping scheme. The simulation results demonstrate that the suggested image encryption method is successful according to the NIST test and has powerful anti-attack, a larger key space as large as 2847, which is prone to multiple attacks, and key sensitivity capabilities. Also, the pixel correlation reached −0.0019, −0.0016, and −0.0069, while the information entropy was at 7.9996. This demonstrates the excellent scrambling effect of the proposed approach, which is capable of greatly improving the color image security performance.

• Ahuja, Bharti; Doriya, Rajesh; Salunke, Sharad; Hashmi, Md. Farukh & Gupta, Aditya (2023). Advanced 5D logistic and DNA encoding for medical images. Imaging Science Journal. ISSN 1368-2199. 71(2), p. 142–160. doi: 10.1080/13682199.2023.2178097. Show summary

  Since patient diagnostic data such as X-rays, MRIs, CT scans, etc. contain very sensitive information about a patient's health, it must be encrypted to prevent unauthorized access. In chaotic-based image encryption methods, the lower-dimensional chaotic maps are insufficient because they can be estimated through some signal estimation technologies; therefore, a higher-order novel 5D logistic chaotic map along with DNA encoding and the Arnold map for medical image encryption is presented to overcome the vital medical breaching that could cost a life. In this work, the advantages of DNA principles and high-dimensional chaotic maps are used to develop an efficient encryption method for the protection of medical images. The strategy is useful due to the fact that the higher order makes the system more complicated and tough to penetrate. The large keyspace up to, anti attack capabilities, and comparison of various standard parameters with different literatures illustrate the algorithm's superiority.

• Gupta, Aditya; Bringsdal, Even; Knausgård, Kristian Muri & Goodwin, Morten (2022). Accurate Wound and Lice Detection in Atlantic Salmon Fish Using a Convolutional Neural Network. Fishes. ISSN 2410-3888. 7(6). doi: 10.3390/fishes7060345. Full text in Research Archive
• Kalhagen, Espen Stausland; Olsen, Ørjan Langøy; Goodwin, Morten & Gupta, Aditya (2022). Hierarchical Object Detection applied to Fish Species. Nordic Machine Intelligence (NMI). ISSN 2703-9196. 2(1), p. 1–15. doi: 10.5617/nmi.9452. Full text in Research Archive Show summary

  Gathering information of aquatic life is often based on timeconsuming methods utilizing video feeds. It would be beneficial to capture more information cost-effectively from video feeds. Video based object detection has an ability to achieve this. Recent research has shown promising results with the use of YOLO for object detection of fish. As underwater conditions can be difficult and thus fish species are hard to discriminate. This study proposes a hierarchical structure-based YOLO Fish algorithm in both the classification and the dataset to gain valuable information. With the use of hierarchical classification and other techniques. YOLO Fish is a state-of-the-art object detector on Nordic fish species, with an mAP of 91.8%. The algorithm has an inference time of 26.4 ms, fast enough to run on real-time video on the high-end GPU Tesla V100.

• Gupta, Aditya; Bringsdal, Even; Salbuvik, Nicole; Knausgård, Kristian Muri & Goodwin, Morten (2022). An Accurate Convolutional Neural Networks Approach to Wound Detection for Farmed Salmon. In Iliadis, Lazaros; Jayne, Chrisina; Tefas, Anastasios & Pimenidis, Elias (Ed.), Engineering Applications of Neural Networks - 23rd International Conference, EAAAI/EANN 2022. Springer Nature. ISSN 978-3-031-08222-1. p. 139–149. doi: 10.1007/978-3-031-082. Show summary

  Fish is one of the most important food sources worldwide and for people in Nordic countries. For this reason, fish has been widely cultivated, but aquacultural fish are severely affected by lice, maturity, wounds, and other harmful factors typically part of agricultural fish, resulting in millions of fish deaths. Unfortunately, diagnosing injuries and wounds in live salmon fish is difficult. However, this study uses image-based machine learning approaches to present a wound detection technique for live farmed salmon fish. As part of this study, we present a new dataset of 3571 photos of injured and non-wounded fish from the Institute of Marine Research’s genuine fish tank. We also propose a Convolutional Neural Network tailored for such wound detection with 20 convolutional and five subsequent dense layers. The model incorporates methods such as dropout, early halting, and Gaussian noise to avoid overfitting. Compared to the established VGG-16 and VGG-19 models, the proposed approaches have a validation accuracy of 96.22%. The model has low 0.0199 and 0.941 false positive and true positive rates, making it a good candidate for accurate live production

View all works in Cristin