SDDGR：用于类别增量目标检测的基于 Stable Diffusion 的深度生成回放
SDDGR: Stable Diffusion-based Deep Generative Replay for Class Incremental Object Detection
AI总结：
针对类别增量目标检测中的灾难性遗忘与多标签复杂性，提出 SDDGR，利用 Stable Diffusion 生成旧类图像，并结合迭代细化、L2 知识蒸馏和伪标签，在 COCO 2017 上取得最优性能。
摘要：
在类别增量学习（CIL）领域，随着生成模型的不断改进，生成回放作为一种缓解灾难性遗忘的方法日益受到重视。然而，其在类别增量目标检测（CIOD）中的应用受到显著限制，主要原因在于涉及多标签场景的复杂性。本文提出了一种用于 CIOD 的新颖方法，称为 stable diffusion deep generative replay（SDDGR）。该方法利用基于扩散的生成模型以及预训练的文本到扩散网络，生成逼真且多样的合成图像。SDDGR 引入迭代细化策略，以生成包含旧类别的高质量图像。此外，我们采用 L2 知识蒸馏技术，以增强模型对合成图像中先前知识的保留。进一步地，该方法还包括对新任务图像中的旧目标进行伪标签标注，防止其被错误分类为背景元素。在 COCO 2017 数据集上的大量实验表明，SDDGR 显著优于现有算法，在多种 CIOD 场景中达到了新的最优水平。源代码将向公众开放。
英文摘要：
In the field of class incremental learning (CIL), generative replay has become increasingly prominent as a method to mitigate the catastrophic forgetting, alongside the continuous improvements in generative models. However, its application in class incremental object detection (CIOD) has been significantly limited, primarily due to the complexities of scenes involving multiple labels. In this paper, we propose a novel approach called stable diffusion deep generative replay (SDDGR) for CIOD. Our method utilizes a diffusion-based generative model with pre-trained text-to-diffusion networks to generate realistic and diverse synthetic images. SDDGR incorporates an iterative refinement strategy to produce high-quality images encompassing old classes. Additionally, we adopt an L2 knowledge distillation technique to improve the retention of prior knowledge in synthetic images. Furthermore, our approach includes pseudo-labeling for old objects within new task images, preventing misclassification as background elements. Extensive experiments on the COCO 2017 dataset demonstrate that SDDGR significantly outperforms existing algorithms, achieving a new state-of-the-art in various CIOD scenarios. The source code will be made available to the public.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
