arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2409.09108cs.LGcs.CVcs.SE
收藏

削减风险:面向深度学习检测系统的可靠持续训练

Trimming the Risk: Towards Reliable Continuous Training for Deep Learning Inspection Systems

Polytechnique Montreal(蒙特利尔理工学院);Sycodal Inc.(Sycodal公司)

Altaf Allah Abbassi, Houssem Ben Braiek, Foutse Khomh, Thomas Reid

AI总结:

针对工业深度学习检测系统持续训练中自标注标签导致性能下降的问题,提出两阶段过滤的鲁棒维护方法,可减少错误自标注数据,提升生产性能且不损害原始验证效果。

摘要:

制造业检测领域正越来越多地依托深度学习(DL)技术,这类检测任务难以通过基于规则的机器视觉算法实现自动化。基于DL的检测系统从标注图像中提取缺陷模式,兼具类人的灵活性与计算机系统的一致性。然而,有限的标注数据集往往无法涵盖所有自然变化,因此需要持续训练(Continuous Training, CT)来利用最新数据定期调整模型。有效的持续训练需要来自原始分布的全新标注样本;否则,自生成的标签可能导致难以察觉的性能下降。为缓解这一风险,我们开发了一种鲁棒的基于持续训练的维护方法,通过两阶段过滤过程筛选可靠数据来更新DL模型。第一阶段过滤掉低置信度预测,因为模型本身就不信任这些预测。第二阶段利用变分自编码器(variational auto-encoders)和直方图生成图像嵌入,以捕捉潜在特征与像素特征,随后拒绝嵌入偏移过大的输入,将其判定为存在错误过度自信的漂移数据。之后,在过滤后的输入数据上对原始DL模型进行微调,同时在最新生产数据与原始数据集的混合集上进行验证。该策略缓解了灾难性遗忘问题,确保模型能有效适应新的运行条件。在针对冰棍棒印刷和玻璃瓶的工业检测系统上,使用关键真实世界数据集开展的评估显示,过滤后仅保留不到9%的错误自标注数据用于微调,使模型在生产数据上的性能最高提升14%,同时不会损害其在原始验证数据上的结果。

英文摘要:

The industry increasingly relies on deep learning (DL) technology for manufacturing inspections, which are challenging to automate with rule-based machine vision algorithms. DL-powered inspection systems derive defect patterns from labeled images, combining human-like agility with the consistency of a computerized system. However, finite labeled datasets often fail to encompass all natural variations necessitating Continuous Training (CT) to regularly adjust their models with recent data. Effective CT requires fresh labeled samples from the original distribution; otherwise, selfgenerated labels can lead to silent performance degradation. To mitigate this risk, we develop a robust CT-based maintenance approach that updates DL models using reliable data selections through a two-stage filtering process. The initial stage filters out low-confidence predictions, as the model inherently discredits them. The second stage uses variational auto-encoders and histograms to generate image embeddings that capture latent and pixel characteristics, then rejects the inputs of substantially shifted embeddings as drifted data with erroneous overconfidence. Then, a fine-tuning of the original DL model is executed on the filtered inputs while validating on a mixture of recent production and original datasets. This strategy mitigates catastrophic forgetting and ensures the model adapts effectively to new operational conditions. Evaluations on industrial inspection systems for popsicle stick prints and glass bottles using critical real-world datasets showed less than 9% of erroneous self-labeled data are retained after filtering and used for fine-tuning, improving model performance on production data by up to 14% without compromising its results on original validation data.

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。

↑