BuildCV

BuildCV Documentation

Our platform provides a visual node-editor interface that simplifies the process of building custom OpenCV algorithms. By connecting pre-built blocks (nodes), users can create powerful image processing pipelines with minimal coding. The platform supports real-time previews, dynamic parameter adjustments, and seamless integration of multiple algorithmic steps.

Getting Started

1. Image URL

  • Purpose: Accepts Image URL as the starting point for the pipeline
  • Inputs: None (initial Node)
  • Output: Original Image

2. Preview Grayscale

  • Purpose: Converts the input image to grayscale.
  • Inputs: Original image.
  • Outputs: Grayscale image.
  • UI Elements: None (conversion is automatic).
  • Preview: Displays the grayscale image.

3. Gaussian Blur

  • Purpose: Applies Gaussian blur to smooth the input image.
  • Inputs: Grayscale or original image.
  • Parameters: Kernel Size (3x3, 5x5, etc.), Sigma (intensity of blur)
  • Outputs: Blurred image.
  • UI Elements: Dropdown for kernel size, Slider or input field for sigma value.
  • Preview: Displays the blurred image.

4. Threshold

  • Purpose: Applies a thresholding operation to the input image.
  • Inputs: Grayscale image.
  • Parameters: Threshold Value, Maximum Value, Type (BINARY, BINARY_INV, TRUNC, TOZERO, TOZERO_INV)
  • Outputs: Thresholded image.
  • UI Elements: Slider or input field for threshold and max values, Dropdown for threshold type.
  • Preview: Displays the thresholded image.

5. Edge Detection

  • Purpose: Detects edges in the input image using selected methods.
  • Inputs: Grayscale or processed image.
  • Parameters: Method (Sobel, Canny), Threshold1 (lower), Threshold2 (upper)
  • Outputs: Edge-detected image.
  • UI Elements: Dropdown for method, Sliders or input fields for thresholds.
  • Preview: Displays the edge-detected image.

6. Output

  • Purpose: Provides the final processed image for download or further use.
  • Inputs: Any processed image.
  • Outputs: None (end of pipeline).
  • UI Elements: Download button.
  • Preview: Displays the final image.

Building a Workflow

Input

  • Drag the Input Image URL node onto the canvas.
  • Enter the image URL or upload a file to initialize the pipeline.

Add Processing Nodes

  • Connect the output of the Input Image URL node to a Preview Grayscale node.
  • Add a Gaussian Blur node and configure parameters:
  • - Kernel size
  • - Sigma
  • Insert a Threshold node to apply a binary threshold.
  • Use an Edge Detection node to highlight edges with methods like Sobel or Canny.

Finalize with Output

  • Connect the last processing node to the Output node.
  • Preview and download the final image.