TFT Image Sensor Detection of Live Bacteria with Deep Learning
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for detecting and classifying bacteria, such as CMOS image sensor-based time-lapse imaging and TFT-based biosensing, require mechanical scanning and cannot differentiate between live and dead bacteria, are time-consuming, and lack sensitivity and specificity, especially in large sample volumes.
Innovation Solution
A TFT-based image sensor system with a large field-of-view (FOV) that uses time-lapse imaging and deep learning to automatically detect and classify bacterial colonies, eliminating the need for mechanical scanning and providing rapid detection and classification of live bacteria.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of time
If CMOS image sensor-based time-lapse imaging is used for early detection and classification of coliform bacteria, then detection time is reduced by more than 12 hours and species classification accuracy exceeds 80%, but mechanical scanning of the Petri dish is required which is time-consuming and requires additional hardware
Solution Approach 1:
The patent merges the image sensor array directly with the Petri dish substrate to create an integrated TFT-based sensor dish. This integration eliminates the need for separate mechanical scanning hardware while maintaining the ability to capture time-lapse images of bacterial colonies across the entire dish surface simultaneously.
Solution Approach 2:
The patent replaces the mechanical scanning system with a stationary TFT image sensor array that captures images optically. Instead of moving parts to scan the dish, the system uses the fixed sensor array to detect bacterial colonies through optical fields, eliminating mechanical complexity while maintaining detection capability.
2Reliability
If traditional culture-based methods are used for detecting E. coli and total coliform bacteria, then detection is accurate and EPA-approved, but the process takes ≥24 hours for final read-out and requires visual recognition by experts
Solution Approach 1:
The patent performs preliminary automated detection and classification of bacterial colonies during the incubation period itself, rather than waiting for the traditional 24-hour read-out. The TFT sensor array continuously monitors colony formation and characteristics in real-time, enabling early detection at 6-10 hours while maintaining EPA-approved accuracy through automated image analysis algorithms.
Solution Approach 2:
The system enables self-service automated detection where the TFT sensor array and deep learning algorithms automatically identify, count, and classify bacterial colonies without requiring expert visual inspection. The system performs its own analysis of colony morphology and patterns to provide reliable detection results.
3Ease of manufacture
If TFT-based biosensing is used to detect pathogens, then the system is low-cost and scalable, but it cannot differentiate between live and dead bacteria and does not provide quantification of CFU concentration
Solution Approach 1:
The patent utilizes color changes in bacterial colonies as they grow and metabolize on the TFT sensor array. Different live bacteria species exhibit distinct color patterns and growth dynamics that can be captured by the image sensor. By analyzing these color changes over time, the system can differentiate between live and dead bacteria and identify specific species, adding measurement precision to the low-cost TFT platform.
Data Source
AI summary
A bacterial colony-forming-unit (CFU) detection system is disclosed that exploits a thin-film-transistor (TFT)-based image sensor array that saves ˜12 hours compared to the Environmental Protection Agency (EPA)-approved methods. A lensfree imaging modality was built using the TFT image sensor with a sample field-of-view of ˜10 cm2. Time-lapse images of bacterial colonies cultured on chromogenic agar plates were automatically collected at 5-minute intervals. Two deep neural networks were used to detect and count the growing colonies and identify their species. When blindly tested with 265 colonies of E. coli and other coliform bacteria (i.e., Citrobacter and Klebsiella pneumoniae), the system reached an average CFU detection rate of 97.3% at 9 hours of incubation and an average recovery rate of 91.6% at ˜12 hours. This TFT-based sensor can be applied to various microbiological detection methods. The imaging field-of-view of this platform can be cost-effectively increased to >100 cm2.