報(bào)告人：汪祥 教授
報(bào)告題目：A Bayesian Deep Prior-Based Quaternion Matrix Completion for Color Image Inpainting
報(bào)告時(shí)間：2025年12月6日（周六）下午3：00
報(bào)告地點(diǎn)：云龍校區(qū)6號(hào)樓304會(huì)議室
主辦單位：數(shù)學(xué)與統(tǒng)計(jì)學(xué)院、數(shù)學(xué)研究院、科學(xué)技術(shù)研究院
報(bào)告人簡(jiǎn)介：
汪祥，博士、教授、博士生導(dǎo)師，現(xiàn)任南昌大學(xué)數(shù)學(xué)與計(jì)算機(jī)學(xué)院副院長(zhǎng)，南昌大學(xué)數(shù)學(xué)一級(jí)學(xué)科博士學(xué)位點(diǎn)和博士后科研流動(dòng)站負(fù)責(zé)人。獲批多個(gè)省級(jí)人才稱號(hào)，擔(dān)任中國(guó)計(jì)算數(shù)學(xué)分會(huì)理事，中國(guó)高等教育學(xué)會(huì)教育數(shù)學(xué)專委會(huì)常務(wù)理事, 國(guó)家天元數(shù)學(xué)東南中心江西基地執(zhí)行主任，國(guó)際知名期刊《Computational and Applied Mathematics》的副主編。主要從事數(shù)值代數(shù)、人工智能與數(shù)據(jù)科學(xué)等領(lǐng)域的研究，在大規(guī)模稀疏特征值問(wèn)題、線性和非線性矩陣方程的數(shù)值求解、譜聚類等方面取得了一些成果。目前主持(含完成)國(guó)家自然科學(xué)基金4項(xiàng)及省部級(jí)項(xiàng)目十幾項(xiàng)。近幾年以第一作者或通訊作者在ACM、JSC、CCP、NLAA等權(quán)威期刊上共發(fā)表SCI收錄論文80多篇。以第一完成人身份獲江西省自然科學(xué)獎(jiǎng)1項(xiàng)和江西省教學(xué)成果獎(jiǎng)3項(xiàng)。
報(bào)告摘要：
Color image inpainting plays an important role in computer vision, which aims to reconstruct missing regions from the available information. Existing quaternion-based deep inpainting methods often struggle to restore both global structure and natural textures, especially when only a single corrupted image is available for training. To address these challenges, we propose BQAE-TV, a novel model that integrates a quaternion fully connected network to capture global features while incorporating total variation regularization to optimize quaternion matrix completion, producing structurally coherent and visually natural images. Furthermore, a Bayesian inference mechanism is employed to regularize the deep image prior and mitigate overfitting. Experiments demonstrate that BQAE-TV outperforms both traditional and state-of the-art methods in terms of visual quality and quantitative metrics, validating its effectiveness and robustness.