Drishti IC is an automated system designed to detect counterfeit integrated circuits (ICs) by analyzing IC markings, package characteristics, and manufacturer-specific parameters. The solution focuses on marking-based verification and supports both offline and online operation, making it suitable for secure and restricted environments.
PS25162 – Counterfeit IC Detection
Counterfeit ICs introduce serious risks such as system failures, security vulnerabilities, and performance degradation. Manual inspection processes are time-consuming, error-prone, and difficult to scale.
Drishti IC addresses this challenge by providing an automated and systematic approach to IC authenticity verification.
- Detect counterfeit ICs using marking and package analysis
- Automate IC verification to reduce manual inspection
- Support offline operation for secure environments
- Ensure reliable verification using trusted references
The development of Drishti IC was guided by:
- Study of IC marking standards and manufacturer documentation
- Analysis of datasheets and trusted reference platforms:
- Digi-Key
- Mouser
- Element14
- Texas Instruments and other OEM websites
- Review of research literature related to counterfeit electronics detection
The system verifies IC authenticity using the following parameters:
- Part Number
- Manufacturer Identity and Logo
- Package Type
- Pin Count
- Package Dimensions
-
Fully Offline Counterfeit Detection
Performs complete marking analysis and parameter verification without internet connectivity, enabling deployment in secure and restricted environments. -
Online Scraping & Datasheet Parsing
Automatically retrieves and structures verified reference images and datasheet parameters from trusted component platforms when internet access is available. -
Multimodal Analysis Engine
Utilizes computer vision, OCR, and feature-matching techniques to validate IC markings, logos, dimensions, and package characteristics.
- Programming Language: Python
- Computer Vision: OpenCV
- Backend: FastAPI
- Desktop Application: Wails + Go
- Vision–Language Model: Qwen 8B (for multimodal understanding and contextual validation)
- Data Sources: Manufacturer datasheets and distributor platforms
Developed by Team Win Diesel as part of Smart India Hackathon.