Faculty Publications

Permanent URI for this communityhttps://idr.nitk.ac.in/handle/123456789/18736

Publications by NITK Faculty

Browse

Author: search.filters.author.Kamath, S.Has files: No

Search Results

Now showing 1 - 5 of 5
  • No Thumbnail Available
    Item
    Machine learning for mobile wound assessment
    (SPIE spie@spie.org, 2018) Kamath, S.; Sirazitdinova, E.; Deserno, T.M.
    Chronic wounds affect millions of people around the world. In particular, elderly persons in home care may develop decubitus. Here, mobile image acquisition and analysis can provide a good assistance. We develop a system for mobile wound capture using mobile devices such as smartphones. The photographs are acquired with the integrated camera of the device and then calibrated and processed to determine the size of various tissues that are present in a wound, i.e., necrotic, sloughy, and granular tissue. The random forest classifier based on various color and texture features is used for that. These features are Sobel, Hessian, membrane projections, variance, mean, median, anisotropic diffusion, and bilateral as well as Kuwahara filters. The resultant probability output is thresholded using the Otsu technique. The similarity between manual ground truth labeling and the classification is measured. The acquired results are compared to those achieved with a basic technique of color thresholding, as well as those produced by the SVM classifier. The fast random forest was found to produce better results. It is also seen to have a superior performance when the method is applied only to the wound regions having the background subtracted. Mean similarity is 0.89, 0.39, and 0.44 for necrotic, sloughy, and granular tissue, respectively. Although the training phase is time consuming, the trained classifier performs fast enough to be implemented on the mobile device. This will allow comprehensive monitoring of skin lesions and wounds. © 2018 SPIE.
  • No Thumbnail Available
    Item
    Modeling and experimental studies on the dynamics of bolted joint structure: Comparison of three vibration-based techniques for structural health monitoring
    (Springer Science and Business Media Deutschland GmbH info@springer-sbm.com, 2020) Deka, A.; Rao, A.; Kamath, S.; Gaurav, A.; Gangadharan, K.V.
    Detection of inadequate tightening in bolted joints is quintessential to ensure structural rigidity and to prevent catastrophic failure. Studies show that 30% of assembly failures occur due to inadequate tightening. In the present study, three vibration-based techniques are presented and compared to detect inadequate tightening of bolted joints. Variation in the damped natural frequency, variation in the damping ratio, and variation in the dynamic joint stiffness are studied with varying tightening torques in the bolted joint. The results show that all the three dynamic parameters vary with the tightness of the bolted joint. Dynamic joint stiffness varies significantly as opposed to the damping ratio and damped natural frequency as tightening torque reduces. In order to verify the results of dynamic stiffness method, ANSYS is used to model and analyze the joint. The experimental setup used to calculate the parameters consists of two Euler–Bernoulli beams connected with single lap bolted joint. © Springer Nature Singapore Pte Ltd 2020.
  • No Thumbnail Available
    Item
    Engagement Analysis of Students in Online Learning Environments
    (Springer Science and Business Media Deutschland GmbH, 2022) Kamath, S.; Singhal, P.; Jeevan, G.; Annappa, B.
    Engagement rate is considered a metric that measures the extent of engagement a particular content is receiving from the audience. In e-learning settings, educators want to observe the level of interest of learners to appropriately modify their courses and make the educational process more effective. In this paper, an ensemble approach is proposed to detect student engagement levels while watching an e-learning video. The ensemble model consists of a deep convolutional neural network (DCNN) for facial expression recognition and a deep recurrent neural network (DRNN) for establishing a relationship between eye-gaze and engagement intensity. OpenFace 2.0 toolbox abilities are leveraged for feature extraction. Experimental results on the test datasets give an accuracy of 55.64% on DAiSEE and an MSE of 0.0598 on Engagement in the Wild Dataset. © 2022, The Author(s), under exclusive license to Springer Nature Switzerland AG.
  • No Thumbnail Available
    Item
    Performance enhancement of HEVC lossless mode using context-based angular and planar intra predictions
    (Springer, 2020) Kamath, S.; Aparna., P.; Antony, A.
    Lossless mode of High-Efficiency Video Coding (HEVC), the state-of-the-art video coding standard, can be used for distortion-free reconstruction of the input data for a wide variety of applications. HEVC relies on the usage of efficient intra prediction strategies to achieve superior compression than its predecessor H.264. A large amount of spatial redundancy exists in almost all video sequences due to coherence, smoothness and the inherent correlation within the neighboring pixels. In this paper, a context-based intra prediction scheme is proposed to minimize this local redundancy by identifying the edges and textures to appropriately modify the prediction strategy at the pixel level, without further increase in the computational complexity. The variability in the sum of absolute differences and local pixel intensity values are chosen to derive the context of the nearby region around the target pixel in the planar and angular intra prediction modes respectively. The experimental results validate the superiority of the proposed method over the HEVC anchor and other state-of-the-art techniques in the literature. © 2020, Springer Science+Business Media, LLC, part of Springer Nature.
  • No Thumbnail Available
    Item
    Leveraging Text Analysis and Machine Learning Models for Suicide Ideation Detection in Social Media Posts
    (Springer, 2025) Lobo, S.R.; Kamath, S.
    The escalating global suicide rates underscore the urgency to leverage social media platforms for mental health discourse. However, discerning suicidal content amidst mental health discussions poses challenges, given the nuanced progression from mental health disorders to suicidal ideation. This study addresses this issue by analyzing text related to suicide ideation within social media posts. Utilizing a dataset featuring text labels for suicide and non-suicide sourced from the SuicideWatch and Depression subreddits on Kaggle, we employ common word and bigram analysis to uncover patterns and relationships. Additionally, emotion analysis is conducted to discern underlying emotional tones in the text corpus. Machine learning and deep learning models are deployed for suicide ideation classification. Support Vector Machines (SVM) coupled with Universal Sentence Encoder (USE) word embedding achieved the highest performance among machine learning models, yielding an accuracy of 94.8% and an F1-score of 0.948. Notably, the deep learning model MentalBERT outperformed others, attaining an accuracy of 97.7% and an F1-score of 0.98. These findings hold promise for assisting professionals in timely identification and provision of mental health support to individuals exhibiting suicidal tendencies. © The Author(s), under exclusive licence to Springer Nature Singapore Pte Ltd. 2025.