TIET Digital Repository

Thapar Institute of Engineering & Technology (TuDR)

Welcome to Thapar Institute of Engineering & Technology Digital Repository (TuDR).

TuDR is the digital asset management system which integrates the intellectual output in the form of research articles, PhD theses, and M.Tech / M.E. theses. TuDR facilitates the sharing and exchange of intellectual output of the university.

TuDR supports the management of scholarly resources of enduring value to Thapar University. Faculty members, students, and research scholars use TuDR services to share their intellectual work with the global academic community.

Download Archive Request Form: (PDF) | (Word)
Submission Guidelines: Download Submission Guidelines

Facilities at Thapar Institute of Engineering & Technology Digital Repository (TuDR):

The users of TuDR can search, download, and browse the collections of documents.
Publish & share electronic documents.
Provide views & comments.
For creating new Communities or Collections, mail to dspace@thapar.edu

Communities in DSpace

Select a community to browse its collections.

Now showing 1 - 5 of 8

Central Library 6
Departments 4950
Research@TIET 59
Persons or groups engaged in research at TIET
Schools 1731
Sponsored Projects 4

Recent Submissions

Development of an Intelligent Model for Spoken Language Identification
(2026-06-29) Kumar, Gaurav; Bhardwaj, Saurabh
Speech is one of the most natural ways for communicating information. In speech signal the language is a medium to convey messages which consists of sounds, words and grammar. As we are moving fast towards globalized society, there is a need to deal with a variety of languages. Spoken Language identification (SLID) is the automatic process to recognize the identity of the language spoken in a speech sample. SLID is an enabling technology that plays an important role in many multilingual speech processing applications, such as spoken language translation, multilingual speech recognition, and spoken document retrieval. It is also a topic of great interest in the areas of intelligence and security for information distillation. Even with advancements in the field, SLID continues to face several challenges. These include similarities in phonetic structures between languages, variations in speaking styles, background disturbances, and differences in speech like pitch, accent, and pronunciation. The present work set out to design and validate improved models for spoken language identification (SLID) by combining bio-inspired optimization with deep learning. The study examined widely used feature-extraction and classification methods, identified acoustic features that best support language discrimination, developed an optimization-driven deep learning framework, and evaluated the proposed models on a benchmark multilingual dataset. The approach followed a steady progression from model design to empirical testing. The first part of the study introduced DBODL MSLIS framework that integrates Dung Beetle Optimization (DBO) with Long Short-Term Memory (LSTM) networks. Speech samples from the IIIT Spoken Language Dataset were processed to extract four key acoustic features: pitch, energy, zero-crossing rate (ZCR), and discrete wavelet transform (DWT) coefficients. The DBO algorithm was used to tune hyperparameters, improve ii convergence, and reduce the likelihood of the model settling in poor local minima. Experiments showed that DBODL-MSLIS provided strong classification performance and generalization, including under noisy conditions. Compared with the existing classifiers, the model recorded notable gains in accuracy, sensitivity, and F-score and maintained stable learning behaviour. In the second phase, a complementary framework named ASLID-GJODL (Automatic Spoken Language Identification using Golden Jackal Optimization with Deep Learning) was developed to further improve reliability and computational efficiency. Here, speech signals were converted into spectrograms, which were then processed by a Squeeze-andExcitation DenseNet (SE-DenseNet). This network adapts its feature emphasis to highlight language-specific cues while reducing the impact of background noise. Model parameters were optimized using the Golden Jackal Optimization (GJO) method, inspired by the cooperative hunting patterns of golden jackals, allowing an effective balance between exploration and convergence. Both proposed frameworks were tested on two benchmark datasets. Their performance was evaluated using confusion matrices, accuracy, precision, sensitivity, specificity, and F1-score. Results were compared with those of conventional machine-learning approaches. Across evaluations, the proposed models consistently demonstrated higher accuracy, better robustness, and improved computational performance relative to existing methods. This research introduces two hybrid architectures and provides insights into the ways in which metaheuristic optimization can improve the performance of deep learning models in the context of speech analysis. The study proved a significant advancement in the development of more efficient SLID systems through the integration of biologically inspired algorithms: Golden Jackal Optimization (GJO) for hyperparameter tuning and Dung Beetle Optimization (DBO) for feature selection. The results clearly indicates that the integration of computational intelligence with deep neural models can proficiently address the diversity and variability inherent in real-world spoken languages.
Perfectionism coping and well-being: A correlational analysis on Indian University Students
(2026-06-26) Khatri, Bhumika; Kaur, Amanpreet
This study investigates the relationship between perfectionism, well-being and coping in Indian university students. Design of this study is cross-sectional, correlation used for analysis with a sample size of 151 participants between the ages of 18 and 30 years from a single university. Standardized instruments were used for data collection including Flett’s Multidimensional Perfectionism Scale, Perceived Stress Scale (PSS-10), Scale of Positive and Negative Experience (SPANE), Satisfaction with Life Scale(SWLS) and Brief COPE. Correlational analysis indicated a positive association between SOP and adaptive coping and SPP and maladaptive coping. It also indicated that SPP is negatively associated with well-being but SOP was not associated to wellbeing.
A Systems Thinking Approach To Road Safety: An Exploratory Study
(2026-06-26) Ahuja, Anandita; Chowdhury, Ipshita
Road traffic fatalities in India exceed 150,000 annually, yet the cognitive mechanisms underlying unsafe interactions between different road user types in heterogeneous mixed traffic remain largely unexamined. This study applies Distributed Situation Awareness theory, grounded in distributed cognition, to investigate how car drivers, two-wheeler riders, and pedestrians structure and create and maintain their situational awareness while navigating shared urban road space in Patiala, Punjab, India. Thirty participants navigated a 16km route while providing concurrent think-aloud verbal protocols. Transcripts were analysed using semantic network construction and word frequency analysis in Python 3, generating structural metrics including density, diameter, and centrality for each group. Network analysis revealed a structural gradient across groups. Car drivers produced a highly integrated network organised around infrastructure cues, two-wheeler riders showed a moderately dense network centred on gap assessment and social negotiation, and pedestrians produced a fragmented network structured around episodic crossing events. Semantic analysis further identified both shared and unique concepts across road user groups, revealing differences in the organisation of awareness. While motorised road users demonstrated overlap in networks, pedestrians exhibited distinct patterns of awareness associated with crossing, gap assessment, and interaction with traffic. The findings confirm that DSA produces coherent results in heterogeneous non-lane-based traffic and extend the framework's empirical base beyond Western contexts.
HETEROGENEOUS CNN-TRANSFORMER ENSEMBLE WITH PROGRESSIVE MULTI-RESOLUTION TRAINING FOR MUSCULOSKELETAL RADIOGRAPH CLASSIFICATION
(2026-06-23) Adity; Kumar, Sumit; Mehta, Rajesh
Musculoskeletal disorders affect one billion people worldwide, which results in large volumes of X-ray radiographs which challenge radiologists and lead to non-trivial interpretation errors. The Current single-model approaches face challenges when learning distinguishing features across seven anatomical regions with different class distributions simultaneously. Body region-specific class imbalance issues make this even more challenging, as the classifiers are typically biased towards the majority classes if no measures are taken during training. The proposed method addresses this by combining complementary CNN and transformer backbones into a heterogeneous ensemble. In this paper, an ablation study on the MURA dataset for fracture detection using EfficientNet-B4-NoisyStudent, ConvNeXt-Small, and Swin Transformer is performed. Comparative analysis demonstrates that feature extraction in the ensembling approach is more robust than in the individual model, even when all three backbones are trained under identical progressive multi-resolution schedules, bone-specific augmentation, and focal-loss-based regularisation. Before feeding into CNN models, the MURA dataset is preprocessed to reduce noise using Gaussian filtering, and images are enhanced using adaptive contrast-limited adaptive histogram equalisation to extract more relevant features and improve classification accuracy. Six ensemble fusion strategies, including weighted soft voting, rank-based averaging, stacking, and Nelder-Mead optimisation, are systematically evaluated and compared. Gradient-based saliency maps interpret the more influential features for model prediction. The experimental results show an accuracy of 84.78% and a sensitivity of 0.9152 across all seven classes of the dataset. The proposed ensemble approach achieves higher classification accuracy than the individual models.
Barriers to Female Leadership in Higher Education
(2026-06-19) Kaur, Jaskirat; Alreja, Sarika
Women are still noticeably underrepresented in senior leadership roles at colleges and universities, even as more women join the academic workforce. Research shows that obstacles like lingering gender stereotypes, being left out of informal circles, a shortage of mentors, the challenge of balancing work and family, and unclear promotion pathways all make it harder for women to rise in university leadership around the world. This study looks at what holds women back from leadership in higher education, using a multi-phase qualitative approach that brings together the experiences and viewpoints of both male leaders and female employees. In the first phase, the researchers spoke with male administrators in senior roles, using both in-depth interviews and a Thematic Apperception Test (TAT). They analyzed the conversations for common themes, such as the belief that leadership is purely merit-based, hidden gender biases, ideas about who "fits" as a leader, internal politics, and hesitance toward women in authority. The TAT also exposed underlying assumptions about women’s abilities, leadership, and their struggle to juggle work and home life. These observations echo recent reviews showing that hidden cultural barriers can persist even when official policies say men and women are equal. Based on the themes identified in Phase 1, an interview questionnaire was created and given to female participants in lower positions at higher education institutions during Phase 2. Their responses showed experiences of limited access to decision-making roles, lack of support, unequal recognition, excessive scrutiny of their work, and self-doubt influenced by social expectations. Recent studies also show that women in middle and lower academic roles face significant challenges in moving toward senior leadership positions. The findings show a clear gap between how male leaders see things and the actual experiences of women. Male participants often described leadership selection as based on skills and not influenced by gender. In contrast, female participants highlighted subtle bias and structural inequality. The study concludes that obstacles to female leadership in higher education are complex and include psychological, personal, and institutional factors. It suggests clear promotion processes, organized mentoring, leadership training programs, initiatives to raise awareness about gender issues, and changes in organizations to make leadership opportunities more accessible for women in academia.