Hybrid LMS-MMSE inverse halftoning technique

Chang, Pao-Chi; Yu, Che-Sheng; Lee, Tien-Hsu

doi:10.1109/83.892446

Cited by 57 publications

(2 citation statements)

References 20 publications

(19 reference statements)

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…This improves convergence speed but with increased computations. Some of the most popular of those methods may include: Modified NLMS [14], [15], Leaky LMS [16], Sign Error and Sign Data LMS [17], [18], Variable step size LMS [19], frequency response shaped LMS [20], Hybrid LMS [21], Absolute Average Error Adjusted Step-Size LMS [2], [22], Filter Proportionate Arctangent framework-based LMS (FP-ALMS) [23], proportionate NLMS (PNLMS) [24] and other algorithms. The use of evolutionary techniques is also suggested to improve convergence of adaptive filters in the denoising operation of medical signals [25].…”

Section: Introductionmentioning

confidence: 99%

Improved LMS Performance for System Identification Application

Mamdoh Abdulatef,

Rami Alomari,

Ahmed Mahmod

2024

IJCDS

View full text Add to dashboard Cite

The adaptive filtering algorithm of Normalized Least Mean Square (NLMS) is known to be highly efficient in terms of requiring less number of iterations compared to the reference Least Mean Square (LMS) method, at the cost of having increased computations. This is because, while a fixed value is assigned for the step size in LMS method, the step size in NLMS is continuously updated to a new value with each iteration. The performance of the gradient descent method of LMS is found to be highly dependent on the step size value assigned at the beginning of the iterations. Throughout the literature, the whole range of LMS step size within which stability is assured is usually suggested and studied, while the selection of the most suitable value of the step size within this range is still not thoroughly studied. In this work, a new method to help specifying the exact value of step size in LMS is proposed. This method is resembled by using the step size value assigned in the first iteration in NLMS method to set the value of the step size in LMS. This is found through the results to be highly effective in approaching NLMS behavior without having to increase computational burden. The performance of the proposed method is evaluated using the Mean Square Error (MSE), Weight Difference (WD) and absolute error after 1000 iterations. Relying on this way to specify the LMS step size can provide simplicity, accuracy and high convergence speed, not only for system identification, but also for many other adaptive filtering applications.

show abstract

Section: Introductionmentioning

confidence: 99%

Improved LMS Performance for System Identification Application

Mamdoh Abdulatef,

Rami Alomari,

Ahmed Mahmod

2024

IJCDS

View full text Add to dashboard Cite

show abstract

“…Chang, Yu e Lee [2] propuseram um método híbrido denominado LMS-MMSE baseado nos filtros LMS (leastmean-square) e MMSE (minimum mean square error) combinados com a técnica look-up-table para reconstruir imagens em níveis de cinza com alta qualidade. Os resultados descritos nesse trabalho são comparados com aqueles conseguidos através da filtragem gaussiana.…”

unclassified

Meio-Tom Inverso Usando Redes Neurais Artificiais

Araújo,

Kim

2005

Anais Do XXII Simpósio Brasileiro De Telecomunicações

View full text Add to dashboard Cite

Resumo− − − − Meio-tom é o processo que converte uma imagem em níveis de cinza G na imagem binária correspondente B, tal que B aparente ser semelhante à G quando vista de uma certa distância. Meio-tom inverso é o processo reverso: é usado para reconstruir a imagem em níveis de cinza G a partir da imagem binária meio-tom B. Nos últimos anos, diversos métodos têm sido propostos para solucionar o problema de meio-tom inverso. Neste artigo, propomos um novo método baseado na aprendizagem por redes neurais artificiais, que denominamos MLP-IH. Aplicamos MLP-IH em imagens meio-tom difusão de erro e imagens meio-tom disparo ordenado. Depois, comparamos os resultados obtidos com os resultados de dois outros métodos de meio-tom inverso: filtragem gaussiana e aprendizagem por árvores de decisão. MLP-IH mostrou ser superior que a árvore de decisão usando janela e conjuntos de treinamento pequenos. Porém, permanece um problema em aberto como aplicar MLP-IH para grandes conjuntos de treinamento. Palavras-Chaves − − − − Meio-tom inverso, aprendizagem de máquina, redes neurais.

show abstract

Inverse halftoning using super‐resolution image processing

Minami

Azuma

Sugie

2011

IEEJ Transactions Elec Engng

View full text Add to dashboard Cite

This paper focuses on the inverse halftoning problem, that is, the problem of reconstructing continuous‐tone images from halftone images of white and black pixels. In general, the problem does not have a unique solution, since halftoning is a many‐to‐one map from continuous‐tone images to binary ones. To this problem, we provide a simple and useful inverse halftoning method, composed of two steps. The first step is to generate several grayscale images from the original halftone image and low‐pass filters. The next step is to reconstruct a continuous‐tone image from the multiple grayscale images by using super‐resolution image processing. This method allows us to obtain good continuous‐tone images which are comparable to the results of the existing methods. The validity of the proposed method is demonstrated by several experiments. © 2011 Institute of Electrical Engineers of Japan. Published by John Wiley & Sons, Inc.

show abstract

Hybrid LMS-MMSE inverse halftoning technique

Cited by 57 publications

References 20 publications

Improved LMS Performance for System Identification Application

Improved LMS Performance for System Identification Application

Meio-Tom Inverso Usando Redes Neurais Artificiais

Inverse halftoning using super‐resolution image processing

Contact Info

Product

Resources

About