{"id":{"repo_id":"ubc","oai_identifier":"oai:circle.library.ubc.ca:2429/92425"},"canonical_url":"https://search.ndltd.org/etd/ubc/oai:circle.library.ubc.ca:2429/92425","repository":{"repo_id":"ubc","name":"University of British Columbia","base_url":"http://circle.library.ubc.ca/oai/request"},"display":{"title":"Visual quality enhancement of low-dose dental cone-beam computed tomography (CBCT) images","abstract":"Cone Beam Computed Tomography (CBCT) has become an essential imaging modality in modern dental diagnostics, offering high-resolution three-dimensional insights critical for procedures such as implant planning and oral surgery. However, the increased radiation dose associated with high-quality CBCT imaging poses significant health risks, particularly for pediatric patients. As a result, low-dose protocols are increasingly adopted, though they introduce substantial noise, reduced resolution, and diminished brightness and contrast—ultimately compromising diagnostic reliability. This thesis presents a novel, modular framework to enhance the visual quality of low-dose CBCT images, composed of three main components: denoising, spatial resolution enhancement, and inverse tone mapping. First, we propose advanced signal processing and deep learning-based denoising techniques, including a speckle-aware filtering pipeline and a modified Selective Frequency Network (SFNet) trained using realistically synthesized high-dose targets. These methods significantly suppress noise while preserving anatomical fidelity. Second, we introduce two spatial resolution enhancement methods: a signal processing-based All-Directional Edge-aware Pixel Interpolation (ADEPT) algorithm, and a lightweight deep learning model called StereoMamba, which leverages contextual cues from adjacent slices for super-resolution. ADEPT achieves high-quality upscaling by incorporating edge-aware pixel weighting and adaptive sharpness control, while StereoMamba offers data-efficient upscaling tailored to CBCT’s structural characteristics. Finally, we develop a novel inverse Tone Mapping Operator (iTMO) designed specifically for grayscale medical imagery. By segmenting the CBCT image into anatomical regions (air, soft tissue, and bone) and applying region-specific tone curves based on perceptual models, the iTMO enhances brightness and contrast without introducing unnatural artifacts. This results in clinically suitable High Dynamic Range (HDR) CBCT images. Extensive subjective and objective evaluations confirm that each component of our framework surpasses leading methods in its class. Beyond enhancing CBCT, the proposed techniques are broadly applicable to other medical imaging modalities where reducing dose or acquisition time is essential. By significantly improving image clarity under low-dose conditions, our framework empowers clinicians to make accurate diagnoses with safer, faster, and more patient-friendly imaging. This work represents a meaningful step toward redefining the balance between radiation exposure and image quality—paving the way for the next generation of diagnostic imaging that is both high-performing and human-centered.","abstract_html":"Cone Beam Computed Tomography (CBCT) has become an essential imaging modality in modern dental diagnostics, offering high-resolution three-dimensional insights critical for procedures such as implant planning and oral surgery. However, the increased radiation dose associated with high-quality CBCT imaging poses significant health risks, particularly for pediatric patients. As a result, low-dose protocols are increasingly adopted, though they introduce substantial noise, reduced resolution, and diminished brightness and contrast—ultimately compromising diagnostic reliability. This thesis presents a novel, modular framework to enhance the visual quality of low-dose CBCT images, composed of three main components: denoising, spatial resolution enhancement, and inverse tone mapping. First, we propose advanced signal processing and deep learning-based denoising techniques, including a speckle-aware filtering pipeline and a modified Selective Frequency Network (SFNet) trained using realistically synthesized high-dose targets. These methods significantly suppress noise while preserving anatomical fidelity. Second, we introduce two spatial resolution enhancement methods: a signal processing-based All-Directional Edge-aware Pixel Interpolation (ADEPT) algorithm, and a lightweight deep learning model called StereoMamba, which leverages contextual cues from adjacent slices for super-resolution. ADEPT achieves high-quality upscaling by incorporating edge-aware pixel weighting and adaptive sharpness control, while StereoMamba offers data-efficient upscaling tailored to CBCT’s structural characteristics. Finally, we develop a novel inverse Tone Mapping Operator (iTMO) designed specifically for grayscale medical imagery. By segmenting the CBCT image into anatomical regions (air, soft tissue, and bone) and applying region-specific tone curves based on perceptual models, the iTMO enhances brightness and contrast without introducing unnatural artifacts. This results in clinically suitable High Dynamic Range (HDR) CBCT images. Extensive subjective and objective evaluations confirm that each component of our framework surpasses leading methods in its class. Beyond enhancing CBCT, the proposed techniques are broadly applicable to other medical imaging modalities where reducing dose or acquisition time is essential. By significantly improving image clarity under low-dose conditions, our framework empowers clinicians to make accurate diagnoses with safer, faster, and more patient-friendly imaging. This work represents a meaningful step toward redefining the balance between radiation exposure and image quality—paving the way for the next generation of diagnostic imaging that is both high-performing and human-centered.","abstract_has_math":false,"creators":["Mirzaei, Simin"],"institution":"University of British Columbia","degree_name":"Doctor of Philosophy - PhD","degree_level":"doctoral","degree_discipline":"Electrical and Computer Engineering","degree_department":null,"school":null,"contributors":["Nasiopoulos, Panos"],"advisors":[],"committee_chairs":[],"committee_members":[],"year":2025,"date_issued":"2025","date_published":"2025","updated_at":"2026-11-19T15:15:03Z","subjects":[],"languages":["eng"],"rights":["Attribution-NonCommercial-NoDerivatives 4.0 International"],"rights_urls":["http://creativecommons.org/licenses/by-nc-nd/4.0/"],"identifier_entries":[],"person_identifier_entries":[]},"links":{"outbound_url":"http://hdl.handle.net/2429/92425","outbound_label":"Handle","outbound_source":"dc:identifier","record_url":"http://hdl.handle.net/2429/92425","record_label":"Repository record","record_kind":"handle","record_source":"dc:identifier","download_url":"http://circle.library.ubc.ca/bitstream/2429/92425/6/ubc_2025_november_mirzaei_simin.pdf","download_label":"Download PDF","download_source":"dc:identifier","access_status":null,"embargo_until":null,"access_note":null},"source_record":{"url":"http://circle.library.ubc.ca/oai/request?verb=GetRecord&metadataPrefix=etdms&identifier=oai%3Acircle.library.ubc.ca%3A2429%2F92425","prefix":"etdms"},"metadata_groups":[{"id":"people","label":"People","entries":[{"key":"dc:contributor","label":"Contributor","values":["Nasiopoulos, Panos"]},{"key":"dc:creator","label":"Author","values":["Mirzaei, Simin"]}]},{"id":"academic_context","label":"Academic Context","entries":[{"key":"dc:date","label":"Dc Date","values":["2025"]},{"key":"dc:publisher","label":"Institution","values":["University of British Columbia"]},{"key":"dc:type","label":"Dc Type","values":["Text"]},{"key":"thesis:degree_discipline","label":"Discipline","values":["Electrical and Computer Engineering"]},{"key":"thesis:degree_level","label":"Degree Level","values":["doctoral"]},{"key":"thesis:degree_name","label":"Degree Name","values":["Doctor of Philosophy - PhD"]},{"key":"thesis:institution_name","label":"Thesis Institution Name","values":["University of British Columbia"]}]},{"id":"language_rights","label":"Language and Rights","entries":[{"key":"dc:language","label":"Dc Language","values":["eng"]},{"key":"dc:rights","label":"Dc Rights","values":["http://creativecommons.org/licenses/by-nc-nd/4.0/","Attribution-NonCommercial-NoDerivatives 4.0 International"]}]},{"id":"identifiers","label":"Identifiers","entries":[{"key":"dc:identifier","label":"Identifier","values":["http://hdl.handle.net/2429/92425","http://circle.library.ubc.ca/bitstream/2429/92425/6/ubc_2025_november_mirzaei_simin.pdf"]}]},{"id":"additional","label":"Additional Metadata","entries":[{"key":"dc:description","label":"Description","values":["Cone Beam Computed Tomography (CBCT) has become an essential imaging modality in modern dental diagnostics, offering high-resolution three-dimensional insights critical for procedures such as implant planning and oral surgery. However, the increased radiation dose associated with high-quality CBCT imaging poses significant health risks, particularly for pediatric patients. As a result, low-dose protocols are increasingly adopted, though they introduce substantial noise, reduced resolution, and diminished brightness and contrast—ultimately compromising diagnostic reliability. This thesis presents a novel, modular framework to enhance the visual quality of low-dose CBCT images, composed of three main components: denoising, spatial resolution enhancement, and inverse tone mapping. First, we propose advanced signal processing and deep learning-based denoising techniques, including a speckle-aware filtering pipeline and a modified Selective Frequency Network (SFNet) trained using realistically synthesized high-dose targets. These methods significantly suppress noise while preserving anatomical fidelity. Second, we introduce two spatial resolution enhancement methods: a signal processing-based All-Directional Edge-aware Pixel Interpolation (ADEPT) algorithm, and a lightweight deep learning model called StereoMamba, which leverages contextual cues from adjacent slices for super-resolution. ADEPT achieves high-quality upscaling by incorporating edge-aware pixel weighting and adaptive sharpness control, while StereoMamba offers data-efficient upscaling tailored to CBCT’s structural characteristics. Finally, we develop a novel inverse Tone Mapping Operator (iTMO) designed specifically for grayscale medical imagery. By segmenting the CBCT image into anatomical regions (air, soft tissue, and bone) and applying region-specific tone curves based on perceptual models, the iTMO enhances brightness and contrast without introducing unnatural artifacts. This results in clinically suitable High Dynamic Range (HDR) CBCT images. Extensive subjective and objective evaluations confirm that each component of our framework surpasses leading methods in its class. Beyond enhancing CBCT, the proposed techniques are broadly applicable to other medical imaging modalities where reducing dose or acquisition time is essential. By significantly improving image clarity under low-dose conditions, our framework empowers clinicians to make accurate diagnoses with safer, faster, and more patient-friendly imaging. This work represents a meaningful step toward redefining the balance between radiation exposure and image quality—paving the way for the next generation of diagnostic imaging that is both high-performing and human-centered."]},{"key":"dc:format","label":"Dc Format","values":["27965733","application/pdf"]},{"key":"dc:title","label":"Title","values":["Visual quality enhancement of low-dose dental cone-beam computed tomography (CBCT) images"]}]}],"canonical_facts":{"dc:contributor":["Nasiopoulos, Panos"],"dc:creator":["Mirzaei, Simin"],"dc:date":["2025"],"dc:description":["Cone Beam Computed Tomography (CBCT) has become an essential imaging modality in modern dental diagnostics, offering high-resolution three-dimensional insights critical for procedures such as implant planning and oral surgery. However, the increased radiation dose associated with high-quality CBCT imaging poses significant health risks, particularly for pediatric patients. As a result, low-dose protocols are increasingly adopted, though they introduce substantial noise, reduced resolution, and diminished brightness and contrast—ultimately compromising diagnostic reliability. This thesis presents a novel, modular framework to enhance the visual quality of low-dose CBCT images, composed of three main components: denoising, spatial resolution enhancement, and inverse tone mapping. First, we propose advanced signal processing and deep learning-based denoising techniques, including a speckle-aware filtering pipeline and a modified Selective Frequency Network (SFNet) trained using realistically synthesized high-dose targets. These methods significantly suppress noise while preserving anatomical fidelity. Second, we introduce two spatial resolution enhancement methods: a signal processing-based All-Directional Edge-aware Pixel Interpolation (ADEPT) algorithm, and a lightweight deep learning model called StereoMamba, which leverages contextual cues from adjacent slices for super-resolution. ADEPT achieves high-quality upscaling by incorporating edge-aware pixel weighting and adaptive sharpness control, while StereoMamba offers data-efficient upscaling tailored to CBCT’s structural characteristics. Finally, we develop a novel inverse Tone Mapping Operator (iTMO) designed specifically for grayscale medical imagery. By segmenting the CBCT image into anatomical regions (air, soft tissue, and bone) and applying region-specific tone curves based on perceptual models, the iTMO enhances brightness and contrast without introducing unnatural artifacts. This results in clinically suitable High Dynamic Range (HDR) CBCT images. Extensive subjective and objective evaluations confirm that each component of our framework surpasses leading methods in its class. Beyond enhancing CBCT, the proposed techniques are broadly applicable to other medical imaging modalities where reducing dose or acquisition time is essential. By significantly improving image clarity under low-dose conditions, our framework empowers clinicians to make accurate diagnoses with safer, faster, and more patient-friendly imaging. This work represents a meaningful step toward redefining the balance between radiation exposure and image quality—paving the way for the next generation of diagnostic imaging that is both high-performing and human-centered."],"dc:format":["27965733","application/pdf"],"dc:identifier":["http://hdl.handle.net/2429/92425","http://circle.library.ubc.ca/bitstream/2429/92425/6/ubc_2025_november_mirzaei_simin.pdf"],"dc:language":["eng"],"dc:publisher":["University of British Columbia"],"dc:rights":["http://creativecommons.org/licenses/by-nc-nd/4.0/","Attribution-NonCommercial-NoDerivatives 4.0 International"],"dc:title":["Visual quality enhancement of low-dose dental cone-beam computed tomography (CBCT) images"],"dc:type":["Text"],"thesis:degree_discipline":["Electrical and Computer Engineering"],"thesis:degree_level":["doctoral"],"thesis:degree_name":["Doctor of Philosophy - PhD"],"thesis:institution_name":["University of British Columbia"]},"updated_at":"2026-11-19T15:15:03Z"}