.png)
Letting Data Speak!
Case Study
Intelligent OCR for Utility Billing Digitization
About the Client
A German utility invoicing company handling various utilities' usage information through paper forms.
Challenge
The company faced several challenges in its manual form processing:
Time-consuming and error-prone manual data entry from paper forms into the central system.
Lack of scalability in processing increasing volumes of forms.
Need to identify different form types after scanning.
Dealing with tilted or rotated scanned forms.
Accurately mapping extracted text to specific form fields.
Requirement for human override capability to correct digitization errors.
Achieving over 90% accuracy while maintaining high availability and performance.
Key Results
Increased form processing throughput by 60%
Reduced manual data entry errors by 85%
Achieved over 90% accuracy in automated form digitization
Solution
JashDS implemented a comprehensive Intelligent OCR solution to automate the utility billing process:
Developed a Computer Vision model for form classification:
Trained on tagged form images to identify the correct form type from scanned images.
Implemented Intelligent OCR capabilities:
Extracted both printed and handwritten text from forms.
Accurately mapped extracted text to corresponding form fields.
Created a structured database for storing extracted data:
Designed to efficiently store and retrieve processed form information.
Built a web interface for human operators:
Allowed side-by-side viewing of original scanned forms and OCR-processed data.
Enabled operators to correct and override OCR results when necessary.
Deployed a secure and scalable infrastructure:
Utilized Google Cloud Platform (GCP) for hosting the solution.
Implemented a secure VPC network with authorized access only.
Designed for dynamic scaling based on system load.
Form1:
Form2:
Form3:
Form4:
Technologies Used
DenseNet161 architecture for Computer Vision
Focal Loss and Cross-Entropy Loss functions
Django web server
PostgreSQL database
Google Cloud Platform (GCP)
Deep Learning models (for OCR and image processing)
Other Case Study Items
AI Model for Retail Shelf Monitoring
JashDS revolutionized retail shelf management for a major grocery chain by developing an AI-powered real-time monitoring system. The solution utilized advanced computer vision techniques and deep learning models to detect out-of-stock and misplaced products, significantly improving inventory accuracy and enhancing the customer shopping experience while reducing manual labor costs.
AI-Driven Candidate Screening Revolution
JashDS revolutionized a company's hiring process by developing a GenAI-powered candidate screener that reduced time-to-hire by 50% and improved hiring outcomes. The solution leverages advanced language models to conduct dynamic, role-specific interviews, automatically generating and adapting questions based on job descriptions and candidate responses.
AI-Powered Resume-Job Matching Engine
JashDS developed an AI-powered resume-job matching engine for a leading Indian job portal, processing over 1M resumes and 50k job descriptions. The solution employed BERT embeddings and HNSW algorithms to create personalized job recommendations for candidates and streamline resume screening for recruiters, significantly enhancing the job search and recruitment processes.