ZeBROD: Zero-Retraining Based Recognition and Object Detection Framework

TL;DR

ZeBROD is a novel framework for object detection that avoids retraining when new products are added, significantly reducing training time and maintaining high accuracy, especially useful in retail scenarios.

Contribution

The paper introduces ZeBROD, a zero-retraining object detection method combining YOLO11n, DeIT, and cosine similarity, addressing catastrophic forgetting and improving efficiency.

Findings

Achieves nearly 3x training time efficiency over classical methods.

Maintains encouraging accuracy for both new and existing products.

Operates with an average inference time of 580 ms per image on edge devices.

Abstract

Object detection constitutes the primary task within the domain of computer vision. It is utilized in numerous domains. Nonetheless, object detection continues to encounter the issue of catastrophic forgetting. The model must be retrained whenever new products are introduced, utilizing not only the new products dataset but also the entirety of the previous dataset. The outcome is obvious: increasing model training expenses and significant time consumption. In numerous sectors, particularly retail checkout, the frequent introduction of new products presents a great challenge. This study introduces Zero-Retraining Based Recognition and Object Detection (ZeBROD), a methodology designed to address the issue of catastrophic forgetting by integrating YOLO11n for object localization with DeIT and Proxy Anchor Loss for feature extraction and metric learning. For classification, we utilize cosine similarity between the embedding features of the target product and those in the Qdrant vector database. In a case study conducted in a retail store with 140 products, the experimental results demonstrate that our proposed framework achieves encouraging accuracy, whether for detecting new or existing products. Furthermore, without retraining, the training duration difference is significant. We achieve almost 3 times the training time efficiency compared to classical object detection approaches. This efficiency escalates as additional new products are added to the product database. The average inference time is 580 ms per image containing multiple products, on an edge device, validating the proposed framework's feasibility for practical use.