Abstract
Effective inventory management is the secret to success in the current high-speed marketing world. Traditional remote inventory management relies on image processing to identify missing products from shelves but mainly contributes to customer privacy concerns. Unlike previous work that focused on identifying missing specific items, this work aims to identify empty areas on shelves without infringing customer identity. In addition to capturing the exact position of empty shelves in a store, it also captures the regions of empty shelves. On top of this, the system has a function that can also detect low levels of inventory and identify the specific items that require replenishment. This is achieved by leveraging top-end technologies such as Optical Character Recognition (OCR) for labeling products, convolutional neural networks (CNNs) to detect stock levels, and a database-driven stock management system to track in real-time and analyze inventories. The system incorporates the pre-trained Faster R-CNN model currently in use for detecting vacant shelves and a hybrid OCR-CNN model to spot item labels and quantities. The system has a replenishment module for creating monthly reports that aggregate shelf occupancy information, low-stock instances, and replenishment activities. The informative reports provide actionable data, trend analysis, and performance metrics, which can be easily presented to management to drive strategic planning. This breakthrough solution offers a privacy-centric, all-in-one approach to shelf monitoring, low stock detection, inventory replenishment, and performance reporting, addressing pressing requirements in the retail sector.
References
Kumar, Keshaw, Saumya Patel, and Rani Astya. "Smart Shelf Monitoring Using YOLO." In 2025 IEEE International Students' Conference on Electrical, Electronics and Computer Science (SCEECS), IEEE, (2025): 1-6.
Pawar, Shital, D. B. Jadhav, Deepali Godse, Rohini Jadhav, and Shruti Thakur. "Vision-Based Empty Shelf Detection in Retail with Real-Time Telegram Notifications for Efficient Restocking."
Arora, Dhruv, Pranav Mittal, Sindhu Hak Gupta, and Richa Sharma. "Enhancing Retail Efficiency: Design and Implementation of Smart Shelf and Trolley Systems Utilizing IoT Technologies." In 2024 First International Conference on Technological Innovations and Advance Computing (TIACOMP), IEEE, (2024): 180-184.
Zubair, Muhammad, SM Sajjad Haider, Ans Rehman, Muhammad Saqib Javed, Aashir Waleed, and Rimsha Chauhdary. "PackMLP: A Patch Based MLP Model for Intelligent Inventory Forecasting and Automation with A-LIQ QR Scanning Integration."
Ayoola, VICTORIA BUKKY, G. E. O. R. G. E. Osam-nunoo, C. H. I. M. A. Umeaku, and BABATUNDE OLUSOLA Awotiwon. "IoT-driven Smart Warehouses with Computer Vision for Enhancing Inventory Accuracy and Reducing Discrepancies in Automated Systems." (2024).
Iqbal, Ayzah Shaheen. "A Maturity Model to Improve Inventory Management Systems in SMEs." PhD diss., Birmingham City University, 2025.
Kumar, Harish. "Driving Retail Efficiency With Computer Vision: Detecting Empty Shelves And Accurate Product Placement Through Planogram Integration." (2023).
Liu, Yang, Dimitra Kalaitzi, Michael Wang, and Christos Papanagnou. "A Machine Learning Approach to Inventory Stockout Prediction." Journal of Digital Economy (2025).
Plakantara, S. P., Athanasia Karakitsiou, and T. Mantzou. "Managing Risks in Smart Warehouses from the Perspective of Industry 4.0." In Disruptive Technologies and Optimization Towards Industry 4.0 Logistics, Cham: Springer International Publishing, (2024): 1-47.
Ojika, Favour Uche, Wilfred Oseremen Owobu, Olumese Anthony Abieba, Oluwafunmilayo Janet Esan, Bright Chibunna Ubamadu, and Andrew Ifesinachi Daraojimba. "The impact of machine learning on image processing: A conceptual model for real-time retail data analysis and model optimization." Unpublished Manuscript (2022).