Current Organizations / Institutes

The organizations/institutes from the scholar's latest article

Previous Organizations / Institutes

The organizations / institutes from the scholar's other articles in the past ten years


Rupinder Saini has a total of 18 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Evaluation of Classifiers for Non-Invasive Heart Rate Measurement Using Video Magnification

    by Rupinder Saini, Pooja Sharma

    As the current world approaches a better lifestyle, medical problems have also increased and gained a lot of attention in the last couple of years. Among others, heart rate detection has become a vital part of remote healthcare monitoring and diagnostic systems. The contactless heart rate measurement is a recent concept that monitors the colour changes in the skin, specifically the facial region, to assess the heart rate of the human. In this paper, the authors have presented an analysis of various machine learning classifiers and independent component analysis methods to support this concept, as the naked eye visualisation cannot detect minor changes in the face. Hence, machine learning and magnification of the frame become a vital step. The paper describes the general procedure of detecting the heart rate by monitoring the forehead of a live object. The training and classification procedure is discussed in detail, and a comparative analysis based on quantitative parameters for different machine learning classifiers is also presented in this paper. The analysis among various classifiers shows that the neural network with the highest accuracy proved to be a better machine learning classifier for heart rate detection work.

  • Open Access

    Firefly algorithm and DNN for improved contactless heart rate measurement from videos

    by Rupinder Saini, Pooja Sharma, Saurabh Kumar, Sapna Juneja, Deepali Gupta, Ghadir Altuwaijri, Gireesh Kumar

    This study significantly improves heart rate analysis by enabling contactless heart measurement through facial video analysis. The system improves feature selection and classification accuracy by utilising a Deep Neural Network architecture with Firefly optimisation, offering a dependable technique for heart rate measurement from facial films. This advancement holds promise for non-invasive and convenient heart rate monitoring in various applications, including healthcare, fitness tracking, and stress management. This study proposes a novel approach integrating Firefly optimization with a Deep Neural Network (DNN) architecture for feature selection and classification tasks. Through comprehensive evaluation against state-of-the-art methodologies, the proposed algorithm demonstrates superior performance across multiple metrics including precision, recall, F-measure, and accuracy. Notably, achieving a precision score of 90.22% and a recall score of 94.46%, the algorithm outperforms existing methodologies, highlighting its efficacy in accurately identifying and classifying instances within the dataset. When compared with recent methods (Li et al., 2023; Kaur, 2022; Su et al., 2023), our method produced up to 2.1% greater precision, 1.5% greater recall, and 2.2% greater F-measure, which indicates a clear advantage for non-contact heart rate classification. Thus, the work grouped the heart rate values into relevant clinically meaningful categories and the performance of classification was measured using classification metrics.


  • Enhancing Remote Healthcare: Non-Contact Heart Rate Detection Using Machine Learning Approaches

    by Rupinder Saini, Pooja Sharma

    The heart rate analysis significantly improved in the proposed work by enabling contactless heart measurement by facial video analysis. The method encompasses deep neural network (DNN) architectures with firefly optimization, which improves classification accuracy and feature selection effectiveness for heart rate prediction from facial images. The proposed algorithm contributes better performance among metrics which includes precision, recall, F-measure, and accuracy.

  • Amalgamating an intelligent variant of the gravitational search algorithm with decision trees for email spam detection

    by Sheetal Kundra, Rama Kant, Sanjiv Kumar Singh, Rupinder Saini, Brijesh Kumar Gupta, Vishnu Sandeep Reddy Manukonda, Vishal Jain

    In today’s interconnected digital world, email remains a primary mode of digital interaction valued for its suitability in official, academic and business communications. However, despite its utility, email faces significant challenges due to the widespread presence of spam in various forms, such as phishing, suspicious attachments and deceptive content. This issue not only affects the efficiency and security of email communication but also poses a barrier to its reliability. Therefore, it is essential to devise effective methods to tackle the escalating count of spam emails. This research work presents an intriguing methodology to combat the persistent problem of email spam. The proposed method, abbreviated as AIGSADT, is an amalgamation of the intelligent variant of the gravitational search algorithm (IGSA) and decision trees (DTs). The machine learning-based DT algorithm is individually inadequate for dealing with the large amount of constructive data on a certain attribute due to its instability and ineffectiveness. The proposed AIGSADT approach integrates the IGSA algorithm, which is effective in handling large amounts of data to detect email spam. This is achieved by constructing decision trees employing gravitational forces as the means of information transfer through mass agents. Here, the intelligent factor of the IGSA algorithm prevents the trapping of GSA agents in local optima by updating their positions based on information provided by the best and worst agents. The performance of the presented AIGSADT approach is analyzed through experiments conducted on various categories available in the Ling spam dataset. These experimental evaluations aim to analyze the significance of different pre-processing modules across different dataset categories. The comparative analysis indicates the supremacy of the proposed approach compared to state-of-the-art methodologies.

  • Evaluation of Sentiment using Deep Learning and Machine Learning using Word Integration Techniques

    by Narinder Kumar, Kiranpreet Kaur, Rupinder Saini, Sanjay Singla, Shilpa Shilpa, Shilpa

    Sentiment analysis involves deciphering contextual sentiments from textual data, which may originate from various sources such as social media posts, comments, reviews, blogs, tweets, and news articles. This process begins with text preprocessing techniques like stop word removal, lemmatization, and stemming, preparing the text for further analysis. Subsequently, the preprocessed text undergoes transformation into vectors via word embedding techniques, facilitating the application of machine learning and neural network methodologies. Among these, Word2vec and Skip-gram have emerged as effective supervised learning techniques, while GLoVe stands out in the unsupervised learning domain. Additionally, Recurrent Neural Networks (RNNs) and Convolution Neural Networks (CNNs) represent the deep learning approaches for analyzing textual data. This paper delves into how these machine learning and deep learning algorithms, applied within word embedding frameworks, can be instrumental in extracting features. These features can then be classified into positive, negative, or neutral sentiments, offering insights into the moral and ethical perspectives of individuals across various sectors, including business, politics, e-commerce, healthcare, and film reviews. Our findings indicate that deep learning methods surpass traditional machine learning in accuracy and effectiveness in predicting sentiments.


  • Open Access

    LDM: A Systematic Review on Lie detection Methodologies

    by Rupinder Saini, Pooja Rani

    In our day-to-day life, Lie detection has a significant concern. We human beings are very much inaccurate while detecting the liars and We believe in what we are told. Lie detection is important in today’s life, because Concealing the information or faking it can sometimes take you to huge problems. In any areas like airport management[2], criminal investigations, counterterrorism, etc this concept has great importance. It is an evergreen challenging and changing topic. This paper presents the common technique which was followed up till now and why it was not considered effective and a review of Robust solutions to detection of deception. People generally do not[3] always believe on what someone says but also try to visualize their facial expressions. While in Robust solution these facial micro-expressions are identified, which are tiny, natural expressions seen on the individual’s face, when they try to conceal or suppress emotions. In addition, the article also provides the year-wise assessment and analysis of research articles published in the area of Lie detection from 2011 to 2022. In the end, our proposed framework for lie detection system is also presented. This paper cover up current issues as well as challenges that could be helpful to resolve in future research works. The review paper closes up by supporting future directions.