This tool implements three different noise reduction algorithms for smoothing data: Rectangular Averaging, Binomial Median Filtering, and Binomial Averaging. It processes data from a list and displays the results in another list.
Click to Expand
Minor bugs fixed.
Explained NoiseReductionKernelWidth and updated algorithm details in README.md.
-
Rectangular Method
- How it works: This method calculates the average of a window of values centered around each data point. The width of the window (kernel width) determines the number of neighboring data points included in the averaging process.
- Principle: By averaging the values within the window, the noise (random fluctuations) is smoothed out, resulting in a clearer, more stable signal.
- Code Implementation:
// Rectangular if (rbtnRect.Checked == true) { for (int i = 0; i < listBox1.Items.Count; i++) { double sum = 0; int count = 0; for (int kernel_index = -NoiseReductionKernelWidth; kernel_index <= NoiseReductionKernelWidth; kernel_index++) { int dataIndex = i + kernel_index; if (dataIndex >= 0 && dataIndex < listBox1.Items.Count) { count++; sum += Convert.ToDouble(listBox1.Items[dataIndex]); } } listBox2.Items.Add(sum / count); lblCnt2.Text = "Count : " + listBox2.Items.Count; } }
-
Binomial Coefficients Method (Median)
- How it works: This method uses the binomial coefficients to weight the data points within the window. The median of the weighted values is then calculated.
- Principle: The median is less sensitive to extreme values (outliers) than the average, making it effective for noise reduction while preserving the overall trend of the data.
- Code Implementation:
// Binomial coefficient (median) else if (rbtnMed.Checked == true) { for (int i = 0; i < listBox1.Items.Count; i++) { List<Tuple<double, int>> weightedValues = new List<Tuple<double, int>>(); for (int kernel_index = -NoiseReductionKernelWidth; kernel_index <= NoiseReductionKernelWidth; kernel_index++) { int dataIndex = i + kernel_index; if (dataIndex >= 0 && dataIndex < listBox1.Items.Count && (NoiseReductionKernelWidth + kernel_index) >= 0 && (NoiseReductionKernelWidth + kernel_index) < binomialCoefficients.Length) { double value = Convert.ToDouble(listBox1.Items[dataIndex]); int weight = binomialCoefficients[Math.Abs(kernel_index)]; for (int w = 0; w < weight; w++) { weightedValues.Add(new Tuple<double, int>(value, weight)); } } } if (weightedValues.Count > 0) { weightedValues.Sort((x, y) => x.Item1.CompareTo(y.Item1)); double weightedMedian; int midIndex = weightedValues.Count / 2; if (weightedValues.Count % 2 == 0) { weightedMedian = (weightedValues[midIndex - 1].Item1 + weightedValues[midIndex].Item1) / 2.0; } else { weightedMedian = weightedValues[midIndex].Item1; } listBox2.Items.Add(weightedMedian); lblCnt2.Text = "Count : " + listBox2.Items.Count; } } }
-
Binomial Coefficients Method (Average)
- How it works: This method uses binomial coefficients to weight the values within the kernel. The weighted average of these values is then calculated.
- Principle: Binomial coefficients give more weight to the central values in the window, which helps to preserve the central trend while smoothing out noise.
- Code Implementation:
// Binomial coefficient (average) else if (rbtnAvg.Checked == true) { binomialCoefficients = AvgCalcBinomialCoefficients(NoiseReductionKernelWidth * 2 + 1); for (int i = 0; i < listBox1.Items.Count; i++) { List<double> values = new List<double>(); int coefficientSum = 0; for (int kernel_index = -NoiseReductionKernelWidth; kernel_index <= NoiseReductionKernelWidth; kernel_index++) { int dataIndex = i + kernel_index; if (dataIndex >= 0 && dataIndex < listBox1.Items.Count) { values.Add(Convert.ToDouble(listBox1.Items[dataIndex]) * binomialCoefficients[NoiseReductionKernelWidth + kernel_index]); coefficientSum += binomialCoefficients[NoiseReductionKernelWidth + kernel_index]; } } if (values.Count > 0) { double sum = values.Sum(); double average = sum / coefficientSum; listBox2.Items.Add(average); lblCnt2.Text = "Count : " + listBox2.Items.Count; } } }
- Calculating Binomial Coefficients
- Principle: Binomial coefficients are derived from Pascal's triangle and represent the coefficients in the expansion of a binomial expression.
- Code Implementation:
private int[] AvgCalcBinomialCoefficients(int windowSize) { int[] coefficients = new int[windowSize]; coefficients[0] = 1; for (int i = 1; i < windowSize; i++) { coefficients[i] = coefficients[i - 1] * (windowSize - i) / i; } return coefficients; }
These algorithms help smooth out noise in data by averaging or filtering based on the chosen method.
- Implemented drag-and-drop functionality to allow users to easily add data to the application.
- Used regular expressions to extract and parse numerical data from various formats.
- Designed and developed a user-friendly interface with interactive elements like buttons and list boxes.
- Provided real-time feedback to the user by updating counts and results dynamically.
- Allowed users to select the noise reduction method and kernel width through the interface.
- Enabled users to calibrate and fine-tune the noise reduction process based on their specific needs.
By implementing these techniques, this project effectively reduces noise from the given data, providing clearer and more reliable results.
This project is licensed under the MIT License. See the LICENSE file for details.
Copyright ⓒ HappyBono 2025. All rights Reserved.