Pharmaceutical Verification Camera System with Learning Algorithm
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Automated mail order pharmacies face bottlenecks in pharmaceutical verification due to the time-consuming and unreliable image capture process of varying pill sizes, colors, and shapes, which leads to inefficiencies in the dispensing system.
Innovation Solution
A learning algorithm-based camera system that stores and applies camera parameters for successful image captures, employing quality checks to evaluate and improve image quality, reducing the need for autofocus and auto-setting features, thereby speeding up the image processing time.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If automated camera systems with autofocus and auto-color balancing features are used to capture images of varying pill sizes, colors, and shapes, then image quality is improved, but image processing time increases significantly
Solution Approach 1:
The system pre-determines and stores optimal camera parameters (focus, white balance, exposure time) for different pill configurations before actual image capture. When a pill is to be imaged, the system retrieves pre-calculated parameters based on pill identification data, eliminating the need for real-time autofocus and auto-color balancing operations.
Solution Approach 2:
The system creates and stores parameter profiles for different pill types, sizes, and configurations. These stored parameter copies are then reused for identical or similar pills, avoiding repeated optimization processes and significantly reducing processing time while maintaining consistent image quality.
2Reliability
If multiple images are captured and processed to ensure quality verification, then reliability of verification is improved, but productivity of dispensing system decreases
Solution Approach 1:
The system replaces manual or complex automated image quality verification processes with a simplified electronic parameter matching system. The camera system automatically compares captured images against stored parameter profiles and verifies quality through digital signal processing, eliminating the need for multiple manual review cycles and physical inspection steps.
Solution Approach 2:
The system implements real-time feedback loops where image capture results are immediately compared against expected parameter ranges. If quality thresholds are met, the process continues automatically; if not, adjustments are made and re-capture occurs. This automated feedback mechanism ensures verification reliability while maintaining high throughput by minimizing re-work cycles.
Data Source
AI summary
A pharmaceutical verification (PV) camera system that captures an image of the contents of a vial on an automated dispensing line is closed. Faster image processing time is achieved by utilizing a learning algorithm that stores camera parameters for a successful image associated with data for a prescription processed on the automated dispensing line. During processing of a prescription order, when the vial contents and availability of stored parameters is confirmed, the stored parameters are transmitted to the camera and an image of the vial contents is captured and stored. When a previously un-encountered or un-trained vial is detected, the camera engages the autofocus feature to capture an image. The learning algorithm evaluates the image based on feedback from one or more metric. Upon agreement with the metric standards, an image is accepted and archived and the camera parameters associated with that vial prescription are stored for later use.