Faculty Publications
Permanent URI for this communityhttps://idr.nitk.ac.in/handle/123456789/18736
Publications by NITK Faculty
Browse
6 results
Search Results
Item Cost-effective real-time aerial surveillance system using edge computing(Springer, 2020) Shahzad Alam, M.; Gupta, S.K.Nowadays there is an emerging need for surveillance in order to maintain the public places more secure and ensure the safety and security of the people. Many government agencies require some autonomous system for surveillance of the large areas which can give them precise and real-time information like number of vehicles, people, and other objects. An aerial surveillance system will be very effective in this scenario and platform like Unmanned Aerial vehicle (UAV) will be very reliable and cost-effective option for this task. To make the system fully autonomous, we require real-time object detection that is computationally complex and time consuming due to the heavy load on the limited processing and payload capacity of low-cost UAV. In this paper, we propose a cost-effective approach for aerial surveillance in which we move the heavy computation tasks to the cloud while keeping limited computation on-board of UAV system using Edge computing technique. Further this will maintain the minimum communication between UAV and the cloud thus proposed system will reduce the network traffic and also delay. Proposed system is based on the state-of-art technique YOLO (You Look Only Once) for real time object detection. © Springer Nature Switzerland AG 2020.Item CVUCAMS: Computer vision based unobtrusive classroom attendance management system(Institute of Electrical and Electronics Engineers Inc., 2018) Gupta, S.K.; Ashwin, T.S.; Guddeti, R.M.One of the major challenges in a smart classroom environment is to develop a computer vision based unobtrusive classroom attendance management system. Traditional classroom environment follows a manual attendance marking system either by calling the student's names or by forwarding an attendance sheet; both interrupts the teaching-learning process and also consume a lot of time. Further, it can be erroneous due to factors such as students' proxy etc. In this paper, we propose an unobtrusive face recognition based smart classroom attendance management system using the high definition rotating camera for capturing the faces of students. The proposed system uses Max-Margin Face Detection (MMFD) technique for the face detection and the model is trained using the Inception-V3 CNN technique for the students' identification. The proposed smart classroom system was tested for a classroom with 20 students at National Institute of Technology Karnataka Surathkal, Mangalore, India and we got the experimental results demonstrate the train and test accuracy of 97.67% and 96.66%, respectively. © 2018 IEEE.Item Optimal Selection of Bands for Hyperspectral Images Using Spectral Clustering(Springer Verlag service@springer.de, 2019) Gupta, V.; Gupta, S.K.; Shukla, D.P.High spectral resolution of hyperspectral images comes hand in hand with high data redundancy (i.e. multiple bands carrying similar information), which further contributes to high computational costs, complexity and data storage. Hence, in this work, we aim at performing dimensionality reduction by selection of non-redundant bands from hyperspectral image of Indian Pines using spectral clustering. We represent the dataset in the form of similarity graphs computed from metrics such as Euclidean, and Tanimoto Similarity using K-Nearest neighbor method. The optimum k for our dataset is identified using methods like Distribution Compactness (DC) algorithm, elbow plot, histogram and visual inspection of the similarity graphs. These methods give us a range for the optimum value of k. The exact value of clusters k is estimated using Silhouette, Calinski-Harbasz, Dunn’s and Davies-Bouldin Index. The value indicated by majority of indices is chosen as value of k. Finally, we have selected the bands closest to the centroids of the clusters, computed by using K-means algorithm. Tanimoto similarity suggests 17 bands out of 220 bands, whereas the Euclidean metric suggests 15 bands for the same. The accuracy of classified image before band selection using support vector machine (SVM) classifier is 76.94% and after band selection is 75.21% & 75.56% for Tanimoto and Euclidean matrices respectively. © 2019, Springer Nature Singapore Pte Ltd.Item Students’ affective content analysis in smart classroom environment using deep learning techniques(Springer New York LLC barbara.b.bertram@gsk.com, 2019) Gupta, S.K.; Ashwin, T.S.; Guddeti, R.M.R.In the era of the smart classroom environment, students’ affective content analysis plays a vital role as it helps to foster the affective states that are beneficial to learning. Some techniques target to improve the learning rate using the students’ affective content analysis in the classroom. In this paper, a novel max margin face detection based method for students’ affective content analysis using their facial expressions is proposed. The affective content analysis includes analyzing four different moods of students’, namely: High Positive Affect, Low Positive Affect, High Negative Affect, and Low Negative Affect. Engagement scores have been calculated based upon the four moods of students as predicted by the proposed method. Further, the classroom engagement analysis is performed by considering the entire classroom as one group and the corresponding group engagement score. Expert feedback and analyzed affect content videos are used as feedback to the faculty member to improve the teaching strategy and hence improving the students’ learning rate. The proposed smart classroom system was tested for more than 100 students of four different Information Technology courses and the corresponding faculty members at National Institute of Technology Karnataka Surathkal, Mangalore, India. The experimental results demonstrate the train and test accuracy of 90.67% and 87.65%, respectively for mood classification. Furthermore, an analysis was performed over incidence, distribution and temporal dynamics of students’ affective states and promising results were obtained. © 2019, Springer Science+Business Media, LLC, part of Springer Nature.Item Automated discontinuity detection and reconstruction in subsurface environment of mars using deep learning: A case study of SHARAD observation(MDPI AG membranes@mdpi.com, 2020) Gupta, V.; Gupta, S.K.; Kim, J.Machine learning (ML) algorithmic developments and improvements in Earth and planetary science are expected to bring enormous benefits for areas such as geospatial database construction, automated geological feature reconstruction, and surface dating. In this study, we aim to develop a deep learning (DL) approach to reconstruct the subsurface discontinuities in the subsurface environment of Mars employing the echoes of the Shallow Subsurface Radar (SHARAD), a sounding radar equipped on the Mars Reconnaissance Orbiter (MRO). Although SHARAD has produced highly valuable information about the Martian subsurface, the interpretation of the radar echo of SHARAD is a challenging task considering the vast stocks of datasets and the noisy signal. Therefore, we introduced a 3D subsurface mapping strategy consisting of radar echo pre-processors and a DL algorithm to automatically detect subsurface discontinuities. The developed components the of DL algorithm were synthesized into a subsurface mapping scheme and applied over a few target areas such as mid-latitude lobate debris aprons (LDAs), polar deposits and shallow icy bodies around the Phoenix landing site. The outcomes of the subsurface discontinuity detection scheme were rigorously validated by computing several quality metrics such as accuracy, recall, Jaccard index, etc. In the context of undergoing development and its output, we expect to automatically trace the shapes of Martian subsurface icy structures with further improvements in the DL algorithm. © 2020 by the authors.Item Fractal-based supervised approach for dimensionality reduction of hyperspectral images(Elsevier Ltd, 2024) Gupta, V.; Gupta, S.K.; Shetty, A.Dimensionality reduction is one of the most challenging and crucial issues apart from data mining, security, and scalability, which have retained much traction due to the ever-growing need to analyze the large volumes of data generated daily. Fractal Dimension (FD) has been successfully used to characterize data sets and has found relevant applications in dimension reduction. This paper presents an application of the FD Reduction (FDR) Algorithm on geospatial hyperspectral data, examining its usefulness for data sets with a relatively high embedding dimension. We examine the algorithm at two levels. First is the conventional FDR approach (unsupervised) at the image level. Alternatively, we propose a pixel-level supervised approach for band reduction based on time-series complexity analysis. Techniques for determining an optimal intrinsic dimension for the dataset using these two techniques are examined. We also develop a parallel GPU-based implementation for the unsupervised image-level FDR algorithm, reducing the run-time by nearly 10 times. Furthermore, both approaches use a support vector machine classifier to compare the classification performance of the original and reduced image obtained. © 2024 Elsevier Ltd