{"id":{"repo_id":"umkc","oai_identifier":"oai:mospace.umsystem.edu:10355/110216"},"canonical_url":"https://search.dev.ndltd.org/etd/umkc/oai:mospace.umsystem.edu:10355/110216","repository":{"repo_id":"umkc","name":"University of Missouri - Kansas City","base_url":"https://mospace.umsystem.edu/oai/request"},"display":{"title":"Complex-valued SAR image compression : an approach for amplitude, phase, and gradient recovery","abstract":"Compressing Synthetic Aperture Radar (SAR) images presents unique challenges due to the high dynamic range and inherent acquisition noise in the amplitude signal, as well as the noise-sensitive and limited information content in the phase signal. Traditional compression methods, such as JPEG and JPEG2000, although widely used, often fail to preserve SAR image quality due to their susceptibility to compression artifacts. The continuous capture of high-resolution raw SAR images over extended periods on drones and Unmanned Aerial Vehicles (UAVs), combined with constraints on computational resources, bandwidth, and onboard storage, further complicates the problem. An effective and efficient compression pipeline is essential for either onboard storage or real-time transmission to ground stations. In this dissertation, we introduce two distinct SAR compression pipelines leveraging deep learning-based approaches tailored for lightweight and end-to-end (E2E) complex-valued SAR image compression. For the lightweight pipeline, we propose a hybrid approach that leverages the Versatile Video Coding (VVC) framework as a core compression engine, combined with a deep learning-based task-specific network, AmpRes and AngRes, which jointly operate in the pixel and transform domains to remove compression artifacts and reconstruct SAR amplitude and phase images. We also introduce the GradRes network to learn gradients for SAR Scale-Invariant Feature Transform (SAR-SIFT), resulting in robust orientation and magnitude estimations. Additionally, we designed a multi-polarization SAR compression based on a distributed and independent encoding approach for a quad-pol SAR image. For the E2E pipeline, we propose a transform-domain-based end-to-end compression network with energy-based latent grouping and multi-rate support for complex-valued SAR images, eliminating the need for off-the-shelf compression engines. Experimental results demonstrate that our proposed approaches outperform state-of-the-art methods in amplitude and phase reconstruction and achieve superior performance in downstream tasks like SAR-SIFT keypoint repeatability.","abstract_html":"Compressing Synthetic Aperture Radar (SAR) images presents unique challenges due to the high dynamic range and inherent acquisition noise in the amplitude signal, as well as the noise-sensitive and limited information content in the phase signal. Traditional compression methods, such as JPEG and JPEG2000, although widely used, often fail to preserve SAR image quality due to their susceptibility to compression artifacts. The continuous capture of high-resolution raw SAR images over extended periods on drones and Unmanned Aerial Vehicles (UAVs), combined with constraints on computational resources, bandwidth, and onboard storage, further complicates the problem. An effective and efficient compression pipeline is essential for either onboard storage or real-time transmission to ground stations. In this dissertation, we introduce two distinct SAR compression pipelines leveraging deep learning-based approaches tailored for lightweight and end-to-end (E2E) complex-valued SAR image compression. For the lightweight pipeline, we propose a hybrid approach that leverages the Versatile Video Coding (VVC) framework as a core compression engine, combined with a deep learning-based task-specific network, AmpRes and AngRes, which jointly operate in the pixel and transform domains to remove compression artifacts and reconstruct SAR amplitude and phase images. We also introduce the GradRes network to learn gradients for SAR Scale-Invariant Feature Transform (SAR-SIFT), resulting in robust orientation and magnitude estimations. Additionally, we designed a multi-polarization SAR compression based on a distributed and independent encoding approach for a quad-pol SAR image. For the E2E pipeline, we propose a transform-domain-based end-to-end compression network with energy-based latent grouping and multi-rate support for complex-valued SAR images, eliminating the need for off-the-shelf compression engines. Experimental results demonstrate that our proposed approaches outperform state-of-the-art methods in amplitude and phase reconstruction and achieve superior performance in downstream tasks like SAR-SIFT keypoint repeatability.","abstract_has_math":false,"creators":["Maharjan, Paras"],"institution":"University of Missouri--Kansas City","degree_name":"Ph.D. (Doctor of Philosophy)","degree_level":"Doctoral","degree_discipline":"Electrical and Electronics Engineering (UMKC)","degree_department":null,"school":null,"contributors":[],"advisors":["Li, Zhu"],"committee_chairs":[],"committee_members":[],"year":2025,"date_issued":"2025","date_published":"2025","updated_at":"2026-07-24T05:18:49Z","subjects":[],"languages":["en_US"],"rights":[],"rights_urls":[],"identifier_entries":[]},"links":{"outbound_url":"https://hdl.handle.net/10355/110216","outbound_label":"Handle","outbound_source":"dc:identifier.uri"},"metadata_groups":[{"id":"people","label":"People","entries":[{"key":"dc:contributor.advisor","label":"Advisor","values":["Li, Zhu"]},{"key":"dc:creator","label":"Author","values":["Maharjan, Paras"]}]},{"id":"academic_context","label":"Academic Context","entries":[{"key":"dc:date.accessioned","label":"Dc Date Accessioned","values":["2026-01-14T20:06:21Z"]},{"key":"dc:date.available","label":"Dc Date Available","values":["2026-01-14T20:06:21Z"]},{"key":"dc:date.issued","label":"Date","values":["2025"]},{"key":"dc:type","label":"Dc Type","values":["Thesis"]},{"key":"thesis:degree_discipline","label":"Discipline","values":["Electrical and Electronics Engineering (UMKC)","Computer Science (UMKC)"]},{"key":"thesis:degree_level","label":"Degree Level","values":["Doctoral"]},{"key":"thesis:degree_name","label":"Degree Name","values":["Ph.D. (Doctor of Philosophy)"]},{"key":"thesis:institution_name","label":"Thesis Institution Name","values":["University of Missouri--Kansas City"]}]},{"id":"language_rights","label":"Language and Rights","entries":[{"key":"dc:language.iso","label":"Language (ISO)","values":["en_US"]}]},{"id":"identifiers","label":"Identifiers","entries":[{"key":"dc:identifier.uri","label":"Identifier URI","values":["https://hdl.handle.net/10355/110216"]}]},{"id":"additional","label":"Additional Metadata","entries":[{"key":"dc:description","label":"Description","values":["Title from PDF of title page viewed February 13, 2026","Dissertation advisor: Zhu Li","Vita","Includes bibliographical references (pages 94-104)","Dissertation (Ph.D.)--School of Computing and Engineering. University of Missouri--Kansas City, 2025"]},{"key":"dc:description.abstract","label":"Abstract","values":["Compressing Synthetic Aperture Radar (SAR) images presents unique challenges due to the high dynamic range and inherent acquisition noise in the amplitude signal, as well as the noise-sensitive and limited information content in the phase signal. Traditional compression methods, such as JPEG and JPEG2000, although widely used, often fail to preserve SAR image quality due to their susceptibility to compression artifacts. The continuous capture of high-resolution raw SAR images over extended periods on drones and Unmanned Aerial Vehicles (UAVs), combined with constraints on computational resources, bandwidth, and onboard storage, further complicates the problem. An effective and efficient compression pipeline is essential for either onboard storage or real-time transmission to ground stations. In this dissertation, we introduce two distinct SAR compression pipelines leveraging deep learning-based approaches tailored for lightweight and end-to-end (E2E) complex-valued SAR image compression. For the lightweight pipeline, we propose a hybrid approach that leverages the Versatile Video Coding (VVC) framework as a core compression engine, combined with a deep learning-based task-specific network, AmpRes and AngRes, which jointly operate in the pixel and transform domains to remove compression artifacts and reconstruct SAR amplitude and phase images. We also introduce the GradRes network to learn gradients for SAR Scale-Invariant Feature Transform (SAR-SIFT), resulting in robust orientation and magnitude estimations. Additionally, we designed a multi-polarization SAR compression based on a distributed and independent encoding approach for a quad-pol SAR image. For the E2E pipeline, we propose a transform-domain-based end-to-end compression network with energy-based latent grouping and multi-rate support for complex-valued SAR images, eliminating the need for off-the-shelf compression engines. Experimental results demonstrate that our proposed approaches outperform state-of-the-art methods in amplitude and phase reconstruction and achieve superior performance in downstream tasks like SAR-SIFT keypoint repeatability."]},{"key":"dc:title","label":"Title","values":["Complex-valued SAR image compression : an approach for amplitude, phase, and gradient recovery"]}]}],"canonical_facts":{"dc:contributor.advisor":["Li, Zhu"],"dc:creator":["Maharjan, Paras"],"dc:date.accessioned":["2026-01-14T20:06:21Z"],"dc:date.available":["2026-01-14T20:06:21Z"],"dc:date.issued":["2025"],"dc:description":["Title from PDF of title page viewed February 13, 2026","Dissertation advisor: Zhu Li","Vita","Includes bibliographical references (pages 94-104)","Dissertation (Ph.D.)--School of Computing and Engineering. University of Missouri--Kansas City, 2025"],"dc:description.abstract":["Compressing Synthetic Aperture Radar (SAR) images presents unique challenges due to the high dynamic range and inherent acquisition noise in the amplitude signal, as well as the noise-sensitive and limited information content in the phase signal. Traditional compression methods, such as JPEG and JPEG2000, although widely used, often fail to preserve SAR image quality due to their susceptibility to compression artifacts. The continuous capture of high-resolution raw SAR images over extended periods on drones and Unmanned Aerial Vehicles (UAVs), combined with constraints on computational resources, bandwidth, and onboard storage, further complicates the problem. An effective and efficient compression pipeline is essential for either onboard storage or real-time transmission to ground stations. In this dissertation, we introduce two distinct SAR compression pipelines leveraging deep learning-based approaches tailored for lightweight and end-to-end (E2E) complex-valued SAR image compression. For the lightweight pipeline, we propose a hybrid approach that leverages the Versatile Video Coding (VVC) framework as a core compression engine, combined with a deep learning-based task-specific network, AmpRes and AngRes, which jointly operate in the pixel and transform domains to remove compression artifacts and reconstruct SAR amplitude and phase images. We also introduce the GradRes network to learn gradients for SAR Scale-Invariant Feature Transform (SAR-SIFT), resulting in robust orientation and magnitude estimations. Additionally, we designed a multi-polarization SAR compression based on a distributed and independent encoding approach for a quad-pol SAR image. For the E2E pipeline, we propose a transform-domain-based end-to-end compression network with energy-based latent grouping and multi-rate support for complex-valued SAR images, eliminating the need for off-the-shelf compression engines. Experimental results demonstrate that our proposed approaches outperform state-of-the-art methods in amplitude and phase reconstruction and achieve superior performance in downstream tasks like SAR-SIFT keypoint repeatability."],"dc:identifier.uri":["https://hdl.handle.net/10355/110216"],"dc:language.iso":["en_US"],"dc:title":["Complex-valued SAR image compression : an approach for amplitude, phase, and gradient recovery"],"dc:type":["Thesis"],"thesis:degree_discipline":["Electrical and Electronics Engineering (UMKC)","Computer Science (UMKC)"],"thesis:degree_level":["Doctoral"],"thesis:degree_name":["Ph.D. (Doctor of Philosophy)"],"thesis:institution_name":["University of Missouri--Kansas City"]},"updated_at":"2026-07-24T05:18:49Z"}