Advanced fraud detection, OCR address extraction, and transaction classification powered by machine learning
Extract structured address and name data from documents with 95%+ accuracy.
Combines rule-based critical checks with ML scoring and TruFor deep learning for comprehensive document forensics.
Automatically categorize bank transactions using ML with 83.73% accuracy across 16 transaction types and 63 categories.
Analyzes metadata, visual artifacts, DPI mismatches, font anomalies, photo fraud, and Benford's Law patterns.
Generates ELA heatmaps and forgery localization maps highlighting tampered regions in documents.
API key authentication and PostgreSQL audit logging.
Contact your administrator to obtain a FORENSICS_API_KEY for authentication.
import requests
url = "https://your-api-domain.com/api/ocr/address"
headers = {"X-API-Key": "your_api_key_here"}
with open("document.pdf", "rb") as f:
files = {"file": f}
response = requests.post(url, headers=headers, files=files)
data = response.json()
print(f"Confidence: {data['confidence']}%")
print(f"Customer Name: {data['extractedData']['MatchResult']['CustomerDetails']['CustomerNames'][0]}")
print(f"Address: {data['extractedData']['MatchResult']['CustomerDetails']['CustomerAddress']}")
const formData = new FormData();
formData.append('file', fileInput.files[0]);
fetch('https://your-api-domain.com/api/ocr/address', {
method: 'POST',
headers: {
'X-API-Key': 'your_api_key_here'
},
body: formData
})
.then(response => response.json())
.then(data => {
console.log(`Confidence: ${data.confidence}%`);
console.log('Extracted Data:', data.extractedData);
})
.catch(error => console.error('Error:', error));
curl -X POST https://your-api-domain.com/api/ocr/address \ -H "X-API-Key: your_api_key_here" \ -F "file=@document.pdf"
Extract structured address and name data from documents (PDFs, images) with confidence scores.
Processing Time: ~12-15 seconds | Accuracy: 95%+
Comprehensive fraud detection combining rule-based checks, ML scoring, and TruFor deep learning analysis.
Processing Time: 2-3 seconds (GPU) | Features: 15+ detection methods
Classify bank transactions into categories using machine learning (TF-IDF + Logistic Regression).
Processing Time: ~6ms (3 transactions) | Accuracy: 83.73% (type), 56.19% (category)