YOLO-IOD: Towards Real Time Incremental Object Detection

TL;DR

YOLO-IOD is a real-time incremental object detection framework built on YOLO, addressing catastrophic forgetting through conflict-aware pseudo-label refinement, importance-based kernel selection, and asymmetric knowledge distillation, with improved performance on realistic benchmarks.

Contribution

The paper introduces YOLO-IOD, a novel real-time incremental object detection method that effectively mitigates knowledge conflicts and catastrophic forgetting in YOLO-based detectors.

Findings

YOLO-IOD outperforms existing methods on standard and realistic benchmarks.

It effectively reduces catastrophic forgetting in incremental detection.

The framework maintains real-time performance while improving accuracy.

Abstract

Current methods for incremental object detection (IOD) primarily rely on Faster R-CNN or DETR series detectors; however, these approaches do not accommodate the real-time YOLO detection frameworks. In this paper, we first identify three primary types of knowledge conflicts that contribute to catastrophic forgetting in YOLO-based incremental detectors: foreground-background confusion, parameter interference, and misaligned knowledge distillation. Subsequently, we introduce YOLO-IOD, a real-time Incremental Object Detection (IOD) framework that is constructed upon the pretrained YOLO-World model, facilitating incremental learning via a stage-wise parameter-efficient fine-tuning process. Specifically, YOLO-IOD encompasses three principal components: 1) Conflict-Aware Pseudo-Label Refinement (CPR), which mitigates the foreground-background confusion by leveraging the confidence levels of pseudo labels and identifying potential objects relevant to future tasks. 2) Importancebased Kernel Selection (IKS), which identifies and updates the pivotal convolution kernels pertinent to the current task during the current learning stage. 3) Cross-Stage Asymmetric Knowledge Distillation (CAKD), which addresses the misaligned knowledge distillation conflict by transmitting the features of the student target detector through the detection heads of both the previous and current teacher detectors, thereby facilitating asymmetric distillation between existing and newly introduced categories. We further introduce LoCo COCO, a more realistic benchmark that eliminates data leakage across stages. Experiments on both conventional and LoCo COCO benchmarks show that YOLO-IOD achieves superior performance with minimal forgetting.