Challenge :
Client faced the need to enhance its operational capabilities for accurately assessing and counting diamonds mounted on jewelry. The traditional methods of manual counting were inefficient and prone to errors, which could undermine the trust and reliability that Client sought to establish in the gemmological community.
Solution :
Computer Vision Algorithm: A deep learning model was developed to analyze high-resolution images of mounted jewelry. This model utilized Convolutional Neural Networks (CNNs) to accurately detect and count diamonds, regardless of the jewelry’s design or lighting conditions.
Flask API: A Flask-based API was created to streamline the process of uploading images and retrieving results. This API allowed clients to easily submit high-resolution images of their jewelry for analysis.
AWS Services:
- The solution leveraged Amazon Web Services (AWS) for data storage and processing.
- Amazon S3 was used to securely store the high-resolution images uploaded by clients.
- AWS Lambda facilitated serverless computing for running the computer vision inference, ensuring scalability and cost-effectiveness.
Outcome:
Increased Efficiency: The automated counting process reduced the time required for diamond assessments from hours to mere minutes.
Accuracy: The system achieved an accuracy rate exceeding 75% in counting diamonds, significantly minimizing human count time, and error.