The ocr object is the result from the ocr model, we use PaddleOCR for this. It contains a list of OCRResult objects.

​
OCR Result

interface OCRResult {
  bbox: BoundingBox;
  text: string;
  confidence: number;
}

Each segment is made form multiple OCRResult objects, and they can be used to identify and annotate the text in the segment. This is useful for tasks such as text extraction, search and more specific highlighting. For segment_type table, each cell is an OCRResult.

Bounding box