Window-Mounted Auction Sticker VIN OCR

The 'Needle in a Haystack' Through Glass

Scanning inventory tags placed on the inside of a vehicle's windshield presents a dual challenge for computer vision. First, there is the extreme text hierarchy discrepancy: auction tags feature massive, bold lot numbers, years, and mileage counts, while the actual VIN is printed in a tiny font near barcodes. Traditional OCR engines default to reading the largest text, outputting useless lot data. Second, shooting through curved automotive glass introduces severe environmental reflections—you often capture mirroring of the sky, clouds, or the inspector's own smartphone, which visually cuts through the small text. For a deeper dive into how we isolate specific data structures, see our core OCR product capabilities.

StructOCR: Built to Filter Noise and Glare

StructOCR doesn't rely on font size or top-to-bottom reading orders. Powered by our flagship Automotive VIN OCR engine, it actively hunts for the 17-character automotive structure. The neural network applies advanced de-glare pre-processing to neutralize the reflections of the sky and the camera lens on the glass. It explicitly ignores the oversized stock numbers, mileage, and barcodes, applying strict ISO 3779 validation to lock onto and extract only the true chassis sequence.

Core Industry Applications

Integration & AI Prompts

Call the endpoint directly, or copy the prompt below into Cursor, Lovable, or v0 to instantly generate your frontend application.

import requests
import base64

# 💰 Save 30%+ vs competitors. Get 20 free credits instantly:
# 👉 https://structocr.com/register

# 1. Prepare Base64 Image
with open("window_auction_tag.jpg", "rb") as image_file:
    base64_image = base64.b64encode(image_file.read()).decode('utf-8')

url = "https://api.structocr.com/v1/vin"
headers = {
    "x-api-key": "YOUR_API_KEY",
    "Content-Type": "application/json"
}
payload = {
    "img": base64_image
}

try:
    print("Scanning Through-Glass Window Sticker...")
    response = requests.post(url, headers=headers, json=payload)
    result = response.json()

    if result.get('success'):
        data = result['data']
        print("✅ Extraction Successful!")
        print(f"VIN:          {data.get('vin')}")
        print(f"Confidence:   {data.get('confidence')}")
        print(f"Carrier Type: {data.get('carrier_type')}")
    else:
        print(f"❌ Extraction Failed: {result.get('error')} - {result.get('message')}")

except Exception as e:
    print(f"An error occurred: {e}")

{
  "success": true,
  "data": {
    "vin": "1G1YB2E46E0123456",
    "confidence": "High",
    "carrier_type": "window_sticker"
  }
}