Add casting from RGB to single channel

[grzanka]

Add an algorithm which could be used to cast RGB images to single channel. There should be couple of options for casting: choose R, G or B or use sqrt(R^2+G^2+B^2).

This way "photos" or "scanned images" could be easily converted into single-channel images suitable for further processing (and displaying using colorbar).

Having this option one coule have following logic to display images on left panel:

• if user uploads RGB image ("picture of cat") it is displayed as normal photo, without colorbar scale.
• if user uploads single channel image ("tiff file with experimental data") it is displayed with proper colorbar scale, helping users to interpret the values.

[grzanka]

Feature Request (by Gemini)

Problem Statement:
Currently, our image processing pipeline lacks a robust and flexible method for converting multi-channel RGB images into single-channel representations. This limitation hinders the seamless integration of photographic or scanned images (e.g., JPEG, PNG, etc.) with scientific data visualization workflows that typically operate on single-channel intensity values and benefit from colormaps for quantitative interpretation.

Proposed Solution: Implement a Configurable RGB to Single-Channel Conversion Algorithm

We propose the addition of a new algorithm capable of converting RGB images to single-channel representations, offering several configurable options to cater to various scientific and display requirements. This feature will enable physicists to efficiently prepare "photos" or "scanned images" for quantitative analysis and visualization alongside experimental data.

Detailed Feature Description:

The core of this feature will be a new image processing function, accessible through both the graphical user interface (GUI) and potentially a scripting API, that performs the RGB to single-channel conversion. The function should offer the following selection of conversion methods:

1. Direct Channel Selection:

   • Red Channel (R): Extract and use only the red channel's intensity values.
   • Green Channel (G): Extract and use only the green channel's intensity values.
   • Blue Channel (B): Extract and use only the blue channel's intensity values.
   • Use Case: Useful when a specific color channel carries the most relevant information or for quick, simple conversions.

2. Luminance-based Conversion (Euclidean Norm):

   • Formula: $I = \sqrt{R^2 + G^2 + B^2}$
   • Rationale: This method calculates the Euclidean norm of the RGB color vector, providing a measure of overall intensity or brightness. This is often a more perceptually accurate representation of "grayscale" than simple averaging, especially in scientific contexts where the magnitude of the color vector is relevant.
   • Implementation Detail: Ensure proper handling of data types to avoid overflow and maintain precision during squaring and square root operations.

3. Luminance-based Conversion (Perceptual Luminance - Optional but Recommended):

   • Formula (e.g., BT.601 standard): $I = 0.299R + 0.587G + 0.114B$
   • Rationale: This formula accounts for the human eye's varying sensitivity to different colors, providing a more perceptually uniform grayscale conversion. While not explicitly requested by the user, this is a standard and highly recommended option for general-purpose image processing where visual interpretation is key.
   • Implementation Detail: This option could be considered as an enhancement in a future iteration or as an additional selectable method.

Integration with Display Logic (Left Panel):

The introduction of this conversion capability will necessitate an update to the image display logic in the left panel to ensure an intuitive and scientifically sound user experience:

• For RGB Images (e.g., "picture of cat"): When a user uploads an image identified as multi-channel (e.g., a JPEG or PNG with 3 or 4 channels), it should be displayed as a normal photographic image. Crucially, no colormap or colorbar scale should be automatically applied to these images, as their interpretation is qualitative rather than quantitative in terms of pixel values.

• For Single-Channel Images (e.g., "tiff file with experimental data"): When a user uploads an image identified as single-channel (e.g., a TIFF file containing grayscale experimental data), it should be displayed with a proper colormap scale and an accompanying colorbar. This visual aid is essential for users to interpret the quantitative values represented by the pixel intensities, facilitating scientific analysis and data comparison.

Benefits:

• Enhanced Data Integration: Seamlessly bridge the gap between photographic/scanned imagery and quantitative scientific data.
• Improved Visualization: Enable the use of colormaps for pseudo-coloring and quantitative interpretation of converted images.
• Streamlined Workflow: Provide a direct and intuitive way for physicists to prepare images for further processing and analysis.
• Increased Flexibility: Offer multiple conversion options to suit different scientific requirements and aesthetic preferences.
• Clarity in Display: Ensure that images are displayed appropriately, with or without colormaps, based on their channel structure and intended interpretation.

Technical Considerations for Programmers:

• Input/Output: The conversion function should accept standard image data structures (e.g., NumPy arrays for Python, or equivalent for other languages) and output a single-channel image of the same data type or a suitable floating-point representation if precision is critical (e.g., for the $\sqrt{R^2 + G^2 + B^2}$ method).
• Performance: The algorithm should be optimized for performance, especially for large images, potentially leveraging vectorized operations (e.g., NumPy operations) or parallel processing where applicable.
• Error Handling: Implement robust error handling for invalid input (e.g., non-RGB images, corrupted data).
• User Interface: Design an intuitive GUI element (e.g., a dropdown menu or radio buttons) for users to select the desired conversion method.
• Metadata Preservation: Consider how relevant image metadata (e.g., EXIF data for photos) should be handled or stripped during the conversion process.
• Scalability: Ensure the solution is scalable to handle images of various sizes and bit depths.

This feature will significantly enhance the utility of our image processing tool for the physics community, providing essential capabilities for data preparation and visualization.