Yield-Adjusted Operation for Convolution Filter Denoising

Abstract Background Segmentation of nuclei in cervical cytology pap smear images is a crucial stage in automated cervical cancer screening. The task itself is challenging due to the presence of cervical cells with spurious edges, overlapping cells, neutrophils, and artifacts. Methods After the initial preprocessing steps of adaptive thresholding, in our approach, the image passes through a convolution filter to filter out some noise. Then, contours from the resultant image are filtered by their distinctive contour properties followed by a nucleus size recovery procedure based on contour average intensity value. Results We evaluate our method on a public (benchmark) dataset collected from ISBI and also a private real dataset. The results show that our algorithm outperforms other state-of-the-art methods in nucleus segmentation on the ISBI dataset with a precision of 0.978 and recall of 0.933. A promising precision of 0.770 and a formidable recall of 0.886 on the private real dataset indicate that our algorithm can effectively detect and segment nuclei on real cervical cytology images. Tuning various parameters, the precision could be increased to as high as 0.949 with an acceptable decrease of recall to 0.759. Our method also managed an Aggregated Jaccard Index of 0.681 outperforming other state-of-the-art methods on the real dataset. Conclusion We have proposed a contour property-based approach for segmentation of nuclei. Our algorithm has several tunable parameters and is flexible enough to adapt to real practical scenarios and requirements.

Download Full-text

Holistic quaternion vector convolution filter for RGB-depth video contour detection

Journal of Electronic Imaging ◽

10.1117/1.jei.26.3.033010 ◽

2017 ◽

Vol 26 (3) ◽

pp. 033010

Author(s):

Chunli Ti ◽

Guodong Xu ◽

Yudong Guan ◽

Yidan Teng ◽

Ye Zhang

Keyword(s):

Contour Detection ◽

Convolution Filter ◽

Depth Video

Download Full-text

EFFICIENT LINE DETECTION METHOD BASED ON 2D CONVOLUTION FILTER

Informatyka Automatyka Pomiary w Gospodarce i Ochronie Środowiska ◽

10.35784/iapgos.2817 ◽

2021 ◽

Vol 11 (4) ◽

pp. 22-27

Author(s):

Paweł Kowalski ◽

Piotr Tojza

Keyword(s):

Detection Method ◽

Computational Cost ◽

New Method ◽

Line Detection ◽

Curve Detection ◽

Real Time Image ◽

Convolution Filter ◽

Straight Lines ◽

Time Image ◽

Local Detection

The article proposes an efficient line detection method using a 2D convolution filter. The proposed method was compared with the Hough transform, the most popular method of straight lines detection. The developed method is suitable for local detection of straight lines with a slope from -45˚ to 45˚. Also, it can be used for curve detection which shape is approximated with the short straight sections. The new method is characterized by a constant computational cost regardless of the number of set pixels. The convolution is performed using the logical conjunction and sum operations. Moreover, design of the developed filter and the method of filtration allows for parallelization. Due to constant computation cost, the new method is suitable for implementation in the hardware structure of real-time image processing systems.

Download Full-text

Convolutional Neural Networks Model in Premature Detection of Melanoma

International Journal of Innovative Technology and Exploring Engineering - Special Issue ◽

10.35940/ijitee.l1002.10812s219 ◽

2019 ◽

Vol 8 (12S2) ◽

pp. 6-11 ◽

Cited By ~ 1

Keyword(s):

Neural Networks ◽

Skin Cancer ◽

Convolutional Neural Networks ◽

Additional Data ◽

Data Set ◽

Non Invasive ◽

Convolution Filter ◽

The University ◽

Improved Accuracy

The objective of this research is provide to the specialists in skin cancer, a premature, rapid and non-invasive diagnosis of melanoma identification, using an image of the lesion, to apply to the treatment of a patient, the method used is the architecture contrast of Convolutional neural networks proposed by Laura Kocobinski of the University of Boston, against our architecture, which reduce the depth of the convolution filter of the last two convolutional layers to obtain maps of more significant characteristics. The performance of the model was reflected in the accuracy during the validation, considering the best result obtained, which is confirmed with the additional data set. The findings found with the application of this base architecture were improved accuracy from 0.79 to 0.83, with 30 epochs, compared to Kocobinski's AlexNet architecture, it was not possible to improve the accuracy of 0.90, however, the complexity of the network played an important role in the results we obtained, which was able to balance and obtain better results without increasing the epochs, the application of our research is very helpful for doctors, since it will allow them to quickly identify if an injury is melanoma or not and consequently treat it efficiently.

Download Full-text

Optimized burst truncation in fast-convolution filter bank based waveform generation

2015 IEEE 16th International Workshop on Signal Processing Advances in Wireless Communications (SPAWC) ◽

10.1109/spawc.2015.7227002 ◽

2015 ◽

Cited By ~ 3

Author(s):

Juha Yli-Kaakinen ◽

Markku Renfors

Keyword(s):

Filter Bank ◽

Fast Convolution ◽

Waveform Generation ◽

Convolution Filter

Download Full-text

A comparison of methods for approximating the vertical gradient of one‐dimensional magnetic field data

Geophysics ◽

10.1190/1.1442221 ◽

1986 ◽

Vol 51 (9) ◽

pp. 1725-1735 ◽

Cited By ~ 5

Author(s):

J. W. Paine

Keyword(s):

Magnetic Field ◽

Fourier Transform ◽

Field Data ◽

Total Error ◽

Vertical Gradient ◽

Magnetic Data ◽

Total Field ◽

One Dimensional ◽

Convolution Filter ◽

Principal Factors

The vertical gradient of a one‐dimensional magnetic field is known to be a useful aid in interpretation of magnetic data. When the vertical gradient is required but has not been measured, it is necessary to approximate the gradient using the available total‐field data. An approximation is possible because a relationship between the total field and the vertical gradient can be established using Fourier analysis. After reviewing the theoretical basis of this relationship, a number of methods for approximating the vertical gradient are derived. These methods fall into two broad categories: methods based on the discrete Fourier transform, and methods based on discrete convolution filters. There are a number of choices necessary in designing such methods, each of which will affect the accuracy of the computed values in differing, and sometimes conflicting, ways. A comparison of the spatial and spectral accuracy of the methods derived here shows that it is possible to construct a filter which maintains a reasonable balance between the various components of the total error. Further, the structure of this filter is such that it is also computationally more efficient than methods based on fast Fourier transform techniques. The spacing and width of the convolution filter are identified as the principal factors which influence the accuracy and efficiency of the method presented here, and recommendations are made on suitable choices for these parameters.

Download Full-text

A General Filter for Stretched-Grid Models: Application in Two-Dimension Polar Geometry

Monthly Weather Review ◽

10.1175/mwr-d-11-00058.1 ◽

2012 ◽

Vol 140 (3) ◽

pp. 919-940

Author(s):

Dorina Surcel ◽

René Laprise

Keyword(s):

Computational Cost ◽

Companion Paper ◽

Polar Coordinates ◽

Limited Area ◽

Intermediate Step ◽

Variable Resolution ◽

Polar Latitude ◽

Area Of Interest ◽

Convolution Filter ◽

Cylindrical Polar Coordinates

Abstract Variable-resolution grids are used in global atmospheric models to improve the representation of regional scales over an area of interest: they have reduced computational cost compared to uniform high-resolution grids, and avoid the nesting issues of limited-area models. To address some concerns associated with the stretching and anisotropy of the variable-resolution computational grid, a general convolution filter operator was developed. The convolution filter that was initially applied in Cartesian geometry in a companion paper is here adapted to cylindrical polar coordinates as an intermediate step toward spherical polar latitude–longitude grids. Both polar grids face the so-called “pole problem” because of the convergence of meridians at the poles. In this work the authors will present some details related to the adaptation of the filter to cylindrical polar coordinates for both uniform as well as stretched grids. The results show that the developed operator is skillful in removing the extraneous fine scales around the pole, with a computational cost smaller than that of common polar filters. The results on a stretched grid for vector and scalar test functions are satisfactory and the filter’s response can be optimized for different types of test function and noise one wishes to remove.

Download Full-text

CONVOLUTION FILTERING OF GRAVITY AND MAGNETIC MAPS

Geophysics ◽

10.1190/1.1439572 ◽

1965 ◽

Vol 30 (2) ◽

pp. 281-283 ◽

Cited By ~ 13

Author(s):

P. Edward Byerly

Keyword(s):

Downward Continuation ◽

Filter Theory ◽

Wavenumber Domain ◽

Gravity And Magnetic ◽

Single Ring ◽

Low Pass ◽

Convolution Filter

This note discusses some applications of convolution filter theory to gravity and magnetic maps. Sampling of the field is equivalent to a multiplication, and the corresponding convolution determines the sampling spectrum. If aliasing is acceptably small a specific filtering multiplication in the wavenumber domain corresponds to a convolution of a set of grid coefficients with gridded map values. The application of this theory to single‐ring residuals, certain vertical derivatives, downward continuation, and low‐pass filtering is discussed.

Download Full-text