https://repository.iiitd.edu.in:443/xmlui The DSpace digital repository system captures, stores, indexes, preserves, and distributes digital research material. Sun, 01 Mar 2026 04:13:26 GMT 2026-03-01T04:13:26Z http://repository.iiitd.edu.in/xmlui/handle/123456789/1818 Accelerated reconstruction of HARDI data and fiber orientation estimation Vaish, Ashutosh; Gupta, Anubha (Advisor); Rajwade, Ajit (Advisor) The human brain has a complex network of nerve fibers, particularly in the white matter regions. This nerve fiber network forms the basis of connectivity within brain regions and to the other parts of the body. Traditional imaging techniques like magnetic resonance imaging (MRI) provide an accurate anatomical picture of the brain but offer limited insights into neural connectivity. This limitation is critical, as some neurological disorders can only be diagnosed by examining the alterations in nerve fiber bundles. To address this challenge, diffusion MRI (dMRI) has emerged as a powerful imaging modality capable of accurately characterizing the orientation of axonal fiber bundles. Diffusion MRI is acquired as a collection of MR volumes (referred to as q-samples). Among dMRI techniques, High Angular Resolution Diffusion Imaging (HARDI) is known to produce better fiber orientation representations than Diffusion Tensor Imaging (DTI) and has more practical computational requirements compared to Diffusion Spectrum Imaging (DSI). Despite its potential, HARDI faces challenges in acquisition and post-processing, requiring innovative solutions. For instance, to achieve the better resolution that HARDI offers over DTI, a large number of q-samples are required, which makes the scanning process slow and prone to motion artifacts. Long scanning times can also be inconvenient for patients. On the other hand, reducing the number of q-samples can reduce reconstruction accuracy. To solve this issue, compressive sensing of HARDI data in k-space and/or q-space presents a potential solution to accelerate the scanning process, where the signal is reconstructed later by exploiting the inherent regularity of the signal. There are two prominent sampling schemes in k-space: ‘Cartesian’ and ‘Radial’, with their pros and cons. The Cartesian sampling scheme is more immune to hardware-based distortions, while radial sampling patterns are more immune to motion-induced artifacts. We propose two methods to reconstruct the compressively acquired measurements from the scanner: MSR-HARDI and TL-HARDI. Both methods primarily focus on acquisition through Cartesian sampling schemes in k-space. The first method, MSR-HARDI, utilizes Multiple Sparsity Regularizers in joint (k − q)-space, allowing higher subsampling ratios that are not feasible with only k-space or only q-space subsampling. Additionally, combining regularizers has been shown to yield improved reconstructions compared to individual regularizers. Building upon this work using fixed sparsifying dictionaries, our second method, TL-HARDI, further explores the application of adaptively learned transforms for the accelerated reconstruction of HARDI. This transform is learned using compressively sensed measurements, eliminating the overhead of selecting data-specific fixed sparsifying dictionaries. Further, since the transform is learned on overlapping patches, it captures local image structure effectively, providing an additional denoising effect to the framework. We also recognize that the radial sampling patterns have less pronounced aliasing due to undersampling compared to Cartesian sampling. These advantages can be leveraged particularly in the acquisition of multidimensional signals like HARDI, with a tremendous scope of acceleration. However, despite these benefits, some hardware issues may lead to k-space samples being acquired along deviated radial trajectories, severely degrading the image reconstruction quality. To address this problem, we have proposed a method called CSR-PERT. In CSR-PERT, we investigated a realistic model of gradient delays, which leads to the measurements being acquired from unknown miscentered radial trajectories. This method proposes a joint framework where these perturbed radial trajectories are estimated and used for reconstructing images from the compressively sensed measurements of MRI and HARDI data. After addressing the acquisition aspect, we also explored another important avenue, i.e., the estimation of fiber orientations from HARDI data. Accurate estimation of the local white matter fiber orientations is required for a reliable neural connectivity analysis. In the absence of ground truth, most previous methods relied on assumptions about the physical models of the diffusion signal. This leads to overly simplistic mathematical models, such as DTI, which fail in regions with multiple fiber crossings or excessively complex models that may fail to converge in real time using conventional optimization techniques. Additionally, previous deep learning (DL) methods also attempted to estimate fiber orientations using orientation distribution functions (ODFs), with possible orientations on a predefined set of directions on the sphere, leading to unavoidable discretization errors. To solve these issues, we proposed two methods: FOREST and DL-MuTE. In FOREST, we estimate the peak fiber orientations directly from the diffusion signal using a branched multi-layer perceptron (MLP). In DL-MuTE, we utilize a multi-tensor model to analyze regions with complex tissue structures, such as crossing fibers. This method presents a DL pipeline that uses a branched architecture to estimate individual tensors of a multi-tensor model, which are then used to infer the underlying fiber orientations within an imaging voxel. The models are trained on synthetic data and evaluated on phantoms with known ground truth. The approach is validated on signals with varying noise levels. Wed, 31 Dec 2025 00:00:00 GMT http://repository.iiitd.edu.in/xmlui/handle/123456789/1818 2025-12-31T00:00:00Z http://repository.iiitd.edu.in/xmlui/handle/123456789/1817 On support and recognition problems for sparse hypergraphs Singh, Karamjeet; Raman, Rajiv (Advisor) A hypergraph H is a pair (V, E), where V is a set of vertices, and E is a collection of subsets of V , called hyperedges. They are used to express complex relations, and they generalize graphs where each element of E is a 2-element subset of V . Hypergraphs are one of the most important combinatorial objects of study in theoretical computer science, and have applications in several domains, including network design, scheduling problems, biology, machine learning, etc. Thus, it is important to study their structural properties. Starting with the work of Zykov [Zyk74], Voloshina and Feinberg [VF84], and John- son and Pollack [JP87], researchers have made several attempts to study the structure of a hypergraph by associating with it an appropriate graph. While their initial attempts were to introduce the planarity of a hypergraph, the notion developed in [VF84; JP87] can be generalized and is now called a support. A support for a hypergraph H = (V, E) is a graph Q = (V, F ) such that for each hyperedge E ∈ E, the induced subgraph Q[E] on the elements of E is connected. With this notion, a hypergraph is considered planar if it admits a support that is a planar graph. The concept of support has practical applications in hypergraph visualization, net- work design, and several optimization problems. Although deciding whether a hyper- graph admits a planar support is NP-hard, identifying sufficient conditions for the existence of such supports, particularly sparse or structured ones, remains a compelling research direction. Most of this thesis delves into the construction of supports for various graph classes. This thesis is divided into three parts. In Part (A), we consider hypergraphs defined by subgraphs of a given host graph. Let G = (V, E) be a graph and H be a collection of subgraphs of G. Then the pair (G, H) naturally defines a hypergraph with vertex set V and a hyperedge V (H) for each H ∈ H. We study support construction in three different settings, depending on whether the host graph G belongs to the class of graphs of (i) bounded genus, (ii) outerplanar, or (iii) bounded treewidth. We gave sufficient conditions that ensure the existence of a support from the same family of graphs as G. The results are extended to dual hypergraphs and to a more general setting- the intersection hypergraphs. We also present a fast algorithm for the construction of a planar support with straight-line embedding when the underlying hypergraph is defined by axis-parallel rectangles and points in R2. Part (B) of the thesis explores the role of supports in solving classical problems such as packing, covering, and coloring problems in hypergraphs. We study these problems for hypergraphs arising from subgraphs of a host graph as well as from geometric regions on orientable surfaces, and present approximation results to the packing and covering problems above. Finally, Part (C) turns to abstract hypergraphs and examines the computational complexity of identifying vertex orderings that forbid fixed patterns. We show NP- hardness of this problem for several vertex orderings, and we deduce implications for the recognition of hypergraphs defined by geometric regions in R2. Tue, 13 Jan 2026 00:00:00 GMT http://repository.iiitd.edu.in/xmlui/handle/123456789/1817 2026-01-13T00:00:00Z http://repository.iiitd.edu.in/xmlui/handle/123456789/1816 Building learner-centered educational experiences in virtual reality Belani, Manshul; Sing, Pushpendra (Advisor); Parnami, Aman (Advisor) Virtual Reality (VR), with its immersive and interactive qualities, is increasingly recognized for its potential to transform education. Unlocking this transformative potential requires the principled design and sustainable adoption of Virtual Reality Learning Environments (VRLEs), ensuring they achieve their pedagogical goals and reach and empower diverse learners at scale. From a design perspective, there is a noticeable gap in the availability of principles, guidelines or frameworks tailored specifically for 3D immersive learning environments. Although the Human–Computer Interaction (HCI) literature offers a rich body of design principles for 2D multimedia learning environments, directly applying these to VR is impractical without thorough investigation, given the unique affordances. Beyond design-related gaps, current research on VRLEs also largely relies on ad hoc, one-off interventions, which underscores the need for systematic approaches that enable sustained adoption and integration into mainstream pedagogy. Therefore, situated at the intersection of HCI, VR, and pedagogy, this dissertation advances the field by addressing the critical challenge of designing effective VRLEs while charting pathways for sustained adoption within mainstream education. Through a combination of empirical evaluations, an integration-focused study, and a comprehensive literature synthesis, this work provides evidence-based insights and frameworks that guide both the design and long-term incorporation of VR in education. As part of this work, empirical studies investigate two core design elements of VRLEs, verbal and spatial representations of learning content, to derive implications for effective design in immersive educational contexts. In parallel, it undertakes an exploration of students’ perceptions, challenges, and barriers related to the sustained use of VRLEs aligned to their curriculum. The dissertation further incorporates a literature-based scoping exercise to identify key design parameters that influence learning effectiveness, in the form of a design space. The findings demonstrate that VR design parameters meaningfully influence cognitive load and user experiences, and that the effectiveness of specific design choices varies across instructional contexts. In addition, by drawing on theoretical frameworks such as Self- Determination Theory, the dissertation outlines key considerations and best practices for integrating and sustaining VR into regular curricula. Furthermore, the identified design space, together with the empirical studies, illustrates how this space can be systematically leveraged to structure investigations that, in turn, inform the development of evidence- based design guidelines for VR learning environments. Building on these insights, the dissertation therefore proposes a foundational frame- work to guide the development of design principles and practical recommendations for VRLEs. In doing so, the thesis aims to contribute to bridging the gap between theory and practice, enhancing learning experiences, and advancing the integration of immersive technologies in educational settings. Mon, 01 Dec 2025 00:00:00 GMT http://repository.iiitd.edu.in/xmlui/handle/123456789/1816 2025-12-01T00:00:00Z http://repository.iiitd.edu.in/xmlui/handle/123456789/1791 Modelling and analysis of UAV-Assisted cellular networks R R, Neetu; Ghatak, Gourab (Advisor); Bohara, Vivek Ashok (Advisor); Srivastava, Anand (Advisor) Emerging as a transformative solution in next-generation wireless networks, unmanned-aerial vehicles (UAVs) provide unprecedented flexibility, rapid deployment, and enhanced connectivity. Their integration into conventional cellular networks presents numerous opportunities, such as dynamic coverage expansion, disaster relief and emergency response, and military and surveillance applications. However, it also brings challenges, including energy constraints, fronthaul and backhaul limitations, and mobility and handover management. This thesis explores three critical aspects of UAV-enabled networks: spectrum management in integrated access and backhaul (IAB) networks, mobility management for seamless handovers, and joint UAV activation control and power optimization in UAV enabled cell-free massive multiple input multiple output (mMIMO) networks under fronthaul capacity limitations. In the first part, we investigate spectrum management in UAV-enabled IAB networks, where the UAVs act as access points and relay data to the core network. We address the challenge of optimally allocating limited spectrum resources between access and backhaul links to maximize network efficiency. Our analysis includes disaster recovery scenarios, where optimal UAV positioning and resource partitioning are derived to sustain user connectivity and maximize throughput. Additionally, in urban environments, we introduce cache-enabled UAVs that reduce reliance on backhaul links, improving content delivery performance. Key metrics such as signal to interference noise ratio (SINR) coverage probability and successful content delivery probability are evaluated. The second focus is on mobility management in UAV-enabled networks with mobile users. Frequent handovers (HOs) due to user mobility present significant challenges to network performance. To address this, we propose a caching-based handover management scheme that reduces handover occurrences by utilizing device caching, thereby enhancing quality of service (QoS). Using spatio-temporal analysis, we assess the scheme’s effectiveness in minimizing handover frequency and ensuring seamless connectivity. Additionally, we examine network reliability by analyzing the conditional success probability (CSP) experienced by users in the presence of blockages. Furthermore, we derive the meta distribution (MD) of SINR and mean local delay (MLD), offering deeper insights into network reliability. The third aspect focuses on the joint optimization of UAV activation and power consumption in UAV-based cell-free mMIMO networks. These networks promise uniform service quality over large areas but are constrained by the limited capacity of wireless fronthaul links connecting UAVs to the central processing unit (CPU). We incorporate functional split options, specifically Options 8 and 7.2, to balance fronthaul capacity, computational complexity, and latency. By formulating and solving a joint placement and power optimization problem, we ensure efficient resource utilization while maintaining fair SINR coverage across users. This thesis provides a structured framework for integrating UAVs into cellular networks, addressing key challenges in spectrum management, mobility, and resource optimization, paving the way for more reliable and efficient UAV-based wireless communication systems. Mon, 01 Dec 2025 00:00:00 GMT http://repository.iiitd.edu.in/xmlui/handle/123456789/1791 2025-12-01T00:00:00Z