Conference Papers
Permanent URI for this collectionhttps://idr.nitk.ac.in/handle/123456789/28506
Browse
5 results
Search Results
Item Empirical Study on Multi Convolutional Layer-based Convolutional Neural Network Classifier for Plant Leaf Disease Detection(Institute of Electrical and Electronics Engineers Inc., 2020) Sunil, C.K.; Jaidhar, C.D.; Patil, N.Recognizing the plant disease automatically in real-time by examining a plant leaf image is highly essential for farmers. This work focuses on an empirical study on Multi Convolutional Layer-based Convolutional Neural Network (MCLCNN) classifier to measure the detection efficacy of MCLCNN on recognizing plant leaf image as being healthy or diseased. To achieve this, a set of experiments were conducted with three distinct plant leaf datasets. Each of the experiments were conducted by setting kernel size of 3× 3 and each experiment was conducted independently with different epochs i.e., 50, 75, 100, 125, and 150. The MCLCNN classifier achieved minimum accuracy of 87.47% with 50 epochs and maximum accuracy of 99.25% with 150 epochs for the Peach plant leaves. © 2020 IEEE.Item CNN-based Soil Fertility Classification with Fertilizer Prescription(Institute of Electrical and Electronics Engineers Inc., 2023) Sujatha, M.; Jaidhar, C.D.Soil fertility plays a vital role in crop growth, and thus, the rapid acquisition of soil fertility levels and applying precise fertilizer is significant for sustainable agricultural development. However, obtaining accurate soil fertility estimates proves difficult due to the traditional practice of laboratory analysis of soil samples. This study proposes fertilizer prescriptions based on the Convolutional Neural Networks (CNNs) classifier results. The soil fertility is classified as HIGH, MEDIUM, or LOW fertile based on the chemical measurements of soil parameters, including EC, pH, OC, P, K, S, Zn, B, Cu, Fe, and Mn. The experiments were carried out by varying kernel size from $3\times 3$ to $7\times 7$ and input grid size from $11\times 11$ to $13\times 13$. The proposed approach outperformed with an Accuracy of 97.24% without oversampling the dataset for kernel size $3\times 3$ and input grid size $11\times 11$. Further, for the dataset oversampled using Synthetic Minority Oversampling (SMOTE) technique, the proposed approach achieved the highest Accuracy of 97.52% for kernel size $3\times 3$ and input grid size $12\times 12$. The study helps in the precise application of fertilizers for specific crops based on classification results. © 2023 IEEE.Item Micro-Moment Classification for Anomalous Power Consumption Detection using 1D CNN(Institute of Electrical and Electronics Engineers Inc., 2023) Nayak, R.; Jaidhar, C.D.Identifying anomalous power consumption is essential in improving energy efficiency in buildings. With the help of sensors and other intelligent systems installed in buildings (including smart homes), identifying anomalous power consumption becomes easy. In this work, 1 Dimensional Convolutional Neural Network (1D CNN)-based classification model is proposed to classify the micro-moments to identify the anomalous power consumption in the presence and absence of the consumer. The SimDataset values are normalized, and each instance with ten features is given as input to the 1D CNN. The robustness of the proposed model is defined by experimenting with varying the hyperparameter to obtain the best performance in the standard performance evaluation metrics. The results depicted that the suggested model outperformed the state-of-the-art, producing an accuracy of 96.4% and a weighted average F1-score of 0.962. © 2023 IEEE.Item Classification of Micro-Moment-Based Anomalous Power Consumption Using Transfer Learning(Institute of Electrical and Electronics Engineers Inc., 2023) Nayak, R.; Jaidhar, C.D.The identification of unusual power usage in buildings is crucial for improving energy efficiency. Using an electrical consumption monitoring system can help with energy conservation by identifying unusual energy consumption patterns. This paper suggests a micro-moment-based methodology for detecting abnormal power use. This study makes use of a benchmark dataset called SimDataset, which is used in most of the micro-moment classification-related works. On the images created from the dataset labeled with two classes and five classes, binary and multi-class classifications have both been used. Transfer learning is used by employing pre-trained CNN models, namely DenseNet121, ResNet50V2, and Xception model. The results depicted that the DenseNet121 model has outperformed all other models by giving the best accuracy of 99% and F1-score of 0.984. © 2023 IEEE.Item Detection and Mitigation of IoT Based DDoS Attack Using Extended MUD Enabled Device Profiling Techniques(Springer Science and Business Media Deutschland GmbH, 2025) Thiruppathi, K.; Jaidhar, C.D.In today’s landscape, the burgeoning Internet of Things (IoT) infrastructure underscores the imperative for implementing top-tier security measures to safeguard the IoT realm. This domain has permeated various sectors, including automotive, smart cities, healthcare, industries, and the power sector. The rising ubiquity of IoT devices has drawn the attention of malicious actors, presenting a significant risk of exploitation in insecure, constrained environments. Among the foremost threats in the IoT domain is the Distributed Denial of Service (DDoS) attack, capable of swiftly devastating entire IoT infrastructures. To address this issue, this work proposes a detection and mitigation model aimed at thwarting DDoS attacks in IoT environments. A hybrid feature selection technique is proposed to identify the most effective features for detecting attacks, and Convolutional Neural Network (CNN) is used to identify the suspicious IoT nodes. Further utilized the extended Manufacturer Usage Description (MUD) enabled device profiling techniques to compare the malicious node profiles with existing benign MUD profiles to find the malicious nodes. Linux IPtables is enabled to efficiently filter DDoS attacks. The proposed work is for detecting IoT DDoS attacks, alongside a self-reliant mitigation strategy to effectively filter these attacks. This strategy aims to minimize the impact of blocking legitimate network traffic from IoT devices. The effectiveness of the proposed hybrid feature selection was evaluated by using the CICIoT2023 dataset. © The Author(s), under exclusive license to Springer Nature Switzerland AG 2025.