Barcode Detection & Recovery System — Shopping Mall (Private)
The Challenge
Shopping malls and retail environments rely heavily on barcode scanning for checkout, inventory, and product tracking, but real-world barcodes are often blurred, damaged, faded, or partially obscured due to wear, poor printing, or handling. Standard barcode scanners fail on these degraded codes, forcing manual entry, slowing checkout lines, and creating inventory/tracking errors.
The Solution
Built a system focused on robust barcode detection, extraction, and decoding even under non-ideal conditions, including: Detection — locating barcodes within images/frames despite blur, damage, low contrast, or partial occlusion Extraction — isolating the barcode region from cluttered or noisy backgrounds Decoding — recovering and decoding the barcode data even from degraded or incomplete codes, using image preprocessing/enhancement techniques (e.g., deblurring, contrast correction) before decoding Designed for real-world retail/shopping mall conditions rather than clean, lab-quality barcode images
The Outcome
Improved barcode read success rate in real-world, imperfect conditions — reducing reliance on manual entry at checkout/inventory points, speeding up scanning workflows, and making barcode-dependent processes more reliable in a high-traffic retail environment.
Technologies
- Machine learning
- AWS
- Deep learning
- python
- flask