Skip to main content

Confidence Score

Overview

The Confidence Score provides a numerical value between 0 and 1 for each word detected during OCR processing. It represents the OCR engine's confidence in accurately recognizing and extracting each word.


How It Works

Word-Level Scoring

Each extracted word is assigned an individual confidence score based on:

  • Text clarity
  • Recognition quality
  • OCR engine heuristics

Independent of Context

  • The confidence score measures the accuracy of word detection only.
  • It does not reflect the correctness of the field or its context within the document.

Scoring Range

  • Values range from 0 (no confidence) to 1 (highest confidence).
  • A score above 0.9 is generally considered accurate and reliable.

Use Cases

Quality Assurance

  • Identify low-confidence words for manual review or correction.
  • Ensure data quality in sensitive or high-accuracy workflows.

Automated Processing

  • Trigger fallback workflows (e.g., reprocess or flag data) when scores fall below a set threshold.

Highlighting Low Confidence Areas

  • Visual tools can display low-score regions to help users spot errors quickly.

Notes

  • Confidence scores do not indicate whether a word fits semantically or structurally in the document.
  • They are strictly based on how accurately the OCR engine recognized the visual text.

Response with Confidence Score

This response includes both the parsed data and a version enriched with confidence scores for each extracted field.


Example Response (JSON)

{
"job_id": 10958,
"job_status": "success",
"parsedData": {
"details": {
"name_of_the_power_company": "MADHYANCHAL VIDYUT VITRAN NIGAM LIMITED",
"customer_id": "0732566000",
"bill_amount": "₹5002",
"arrears": "",
"bill_issue_date": "20-03-2025",
"customer_address": "AKHLAK NAGKATRI AKHLAK NAGAR UNNAO UNNAO UP IND"
}
},
"parsedDataWithConfScore": {
"details": {
"name_of_the_power_company": {
"value": "MADHYANCHAL VIDYUT VITRAN NIGAM LIMITED",
"confidence": 98.98
},
"customer_id": {
"value": "0732566000",
"confidence": 97.64
},
"bill_amount": {
"value": "₹5002",
"confidence": 0
},
"arrears": {
"value": "",
"confidence": 0
},
"bill_issue_date": {
"value": "20-03-2025",
"confidence": 0
},
"customer_address": {
"value": "AKHLAK NAGKATRI AKHLAK NAGAR UNNAO UNNAO UP IND",
"confidence": 98.85
}
}
},
"totalAttempts": 1,
"total_pages": 2,
"initiated_at": 1744026828552
}