@techreport{oai:ipsj.ixsq.nii.ac.jp:00174335,
 author = {松岡, 丈平 and 古賀, 崇了 and 末竹, 規哲 and 内野, 英治 and Jyohei, Matsuoka and Takanori, Koga and Noriaki, Suetake and Eiji, Uchino},
 issue = {3},
 month = {Aug},
 note = {本報告では，原信号の構造保存性に優れたランダム値インパルス性雑音除去手法を提案する．構造保存性に優れたランダム値インパルス性雑音除去手法に Non-local median filter がある．Non-local 処理は，着目画素を中心とするブロックと構造が類似したブロックの情報を用いて雑音除去を行う．その結果，原信号の構造を崩すことなく，積極的に雑音除去を行うことができる． しかし，Non-local median filter では，雑音発生確率が高い場合，Non-local 処理の性能は極端に低下するという問題がある． この原因として，雑音の影響により，構造の類似度を正しく評価できないことが考えられる．そこで，本研究では原信号らしさにより重み付けされた類似度を用いる新たな Non-local median filter を提案する．実験を通して，提案手法の有効性を示す．, As one of the random-noise removal methods, there is a non-local median filter. Non-local median filter is excelled in not only noise removal but also detail preserving. In the non-local processing, an output of the filter is calculated from pixels in blocks which are similar to the block centered at a pixel of interest. As a result, aggressive noise removal is conducted without destroying detailed structures in an original image. However, the performance of the non-local median filter decrease enormously when the input image is highly corrupted by the noise. To cope with the problem, we propose an improved non-local median filter which is robust to the high level of corruption by introducing a new similarity measure considering possibility of being original signal. Through the experiments, we verify the effectiveness of the proposed method.},
 title = {重み付き非局所的メディアンフィルタによるランダム値インパルス性雑音除去},
 year = {2016}
}