HSOCRObservation

A single region of text recognized in an image.

Instances are delivered inside the observations array of an HSOCRResult. Each observation represents a discrete text run found in the source image, along with a confidence score and a normalized bounding box.

Bounding-box coordinates use a normalized top-left-origin system: (0, 0) is the top-left corner of the image and (1, 1) is the bottom-right. This matches the convention used by most image-processing tools and differs from Vision's internal bottom-left-origin system (the conversion is automatic).

• Example:

const result = await hs.ocr.recognizeText('/tmp/image.png')
result.observations.forEach(obs => {
    const pct = (obs.confidence * 100).toFixed(0)
    console.log(obs.text + ' (' + pct + '% confidence)')
    const b = obs.bounds
    console.log('  region: x=' + b.x.toFixed(3) + ' y=' + b.y.toFixed(3)
                + ' w=' + b.w.toFixed(3) + ' h=' + b.h.toFixed(3))
})