Struct ImageProcessor 

pub struct ImageProcessor;

Processor for image binarization operations

Provides functionaility to convert images to binary format using Otsu’s thresholding method with parallel processing.

Implementations

impl ImageProcessor

pub fn process_image( image_path: &str, processing_method: ImageProcessingMethod, ) -> Result<ImageBuffer<Luma<u16>, Vec<u16>>, Box<dyn Error>>

Loads an image, converts it to grayscale, and binarizes it using Otsu’s method

Arguments

• image_path - Path to the input image file

Returns

• Ok(ImageBuffer): Binary image buffer with white pixels (255) where original > threshold, black(0) otherwise
• Err: If image loading fails or buffer creation fails

Process

1. Load image from path
2. Convert to 16-bit grayscale
3. Calculate Otsu threshold
4. Apply threshold in parallel to create binary image
5. Return new image buffer

Example

let binary_img = ImageProcessor::process_image("input.png")?;
binary_img.save("binary.png")?;

pub fn calculate_otsu_threshold(img: &ImageBuffer<Luma<u16>, Vec<u16>>) -> u16

Calculates optimal threshold for binarization using Otsu’s method

Arguments

• img - 16-bit grayscale image reference

Returns

Optimal threshold value (0-65535) that maximizes inter-class variance

Process

1. Build 16-bit histogram (0-65535)
2. Iterate all possible thresholds:
   • Split histogram into foreground/background
   • Calculate class weights and means
   • Compute between-class variance
3. Return threshold with maximum variance

pub fn calculate_kapur_threshold(img: &ImageBuffer<Luma<u16>, Vec<u16>>) -> u16

pub fn calculate_sauvola_threshold( img: &ImageBuffer<Luma<u16>, Vec<u16>>, ) -> u16

pub fn calculate_bernsen_threshold( img: &ImageBuffer<Luma<u16>, Vec<u16>>, ) -> u16

pub fn calculate_wolf_threshold(img: &ImageBuffer<Luma<u16>, Vec<u16>>) -> u16