Struct LayoutPostProcess 

pub struct LayoutPostProcess { /* private fields */ }

Layout detection post-processor for models like PicoDet and RT-DETR.

This processor converts model predictions into bounding boxes with class labels and confidence scores for document layout elements.

Implementations

impl LayoutPostProcess

pub fn new( num_classes: usize, score_threshold: f32, nms_threshold: f32, max_detections: usize, model_type: String, ) -> LayoutPostProcess

Creates a new layout detection post-processor.

pub fn apply( &self, predictions: &ArrayBase<OwnedRepr<f32>, Dim<[usize; 4]>>, img_shapes: Vec<ImageScaleInfo>, ) -> (Vec<Vec<BoundingBox>>, Vec<Vec<usize>>, Vec<Vec<f32>>)

Applies post-processing to layout detection model predictions.

Arguments

• predictions - Model output tensor [batch, num_boxes, 4 + num_classes]
• img_shapes - Original image dimensions for each image in batch

Returns

Tuple of (bounding_boxes, class_ids, scores) for each image in batch

Trait Implementations

impl Clone for LayoutPostProcess

fn clone(&self) -> LayoutPostProcess

Returns a duplicate of the value.

fn clone_from(&mut self, source: &Self)

Performs copy-assignment from source.

impl Debug for LayoutPostProcess

fn fmt(&self, f: &mut Formatter<'_>) -> Result<(), Error>

Formats the value using the given formatter.

impl Default for LayoutPostProcess

fn default() -> LayoutPostProcess

Returns the “default value” for a type.