在实时物体检测中重新审视分布外检测:从基准陷到新的缓解范式
IEEE transactions on pattern analysis and machine intelligence
|January 6, 2026
概括
分布之外的输入挑战了深度学习模型. 这项研究揭示了基准缺陷,并引入了一种训练时间方法,以减少对象检测中的幻觉错误,显著提高模型的稳定性.
科学领域:
- 计算机科学 计算机科学
- 人工智能的人工智能
- 机器学习 机器学习
背景情况:
- 深度学习模型与分布外 (OoD) 输入扎,导致对错误对象的过度自信预测.
- 在对象检测中OOD检测的现有方法通过专注于后期评分调整提供了有限的改进.
研究的目的:
- 为解决对象检测模型的OoD检测中的关键被忽视的问题.
- 提出一种新的培训时间减缓策略,以提高对OOD输入的模型稳定性.
主要方法:
- 分析了广泛使用的评估基准,确定了扭曲绩效指标的重大数据质量问题.
- 引入了一种训练时间范式,涉及精细调探测器与合成的OoD数据集,模仿分布式对象.
- 开发了一种方法来抑制OoD对象的对象性,创建一个更具防御性的决策边界.
主要成果:
- 发现OOD测试套件中高达13%的对象被错误标记,影响了现有方法的评估.
- 在BDD-100K数据集上,YOLO模型的幻觉错误减少了91%.
- 在各种物体检测架构 (YOLO,更快的R-CNN,RT-DETR) 中展示了泛化和少数镜头适应.
结论:
- 当前的OoD检测基准有根本的缺陷,这会夸大性能要求.
- 一种独立于外部检测器的训练时间缓解方法有效地减少了OOD诱导的幻觉.
- 拟议的方法提供了一个原则性和广泛适用的解决方案,以提高物体检测器的可靠性.
相关概念视频
Difference from Background: Limit of Detection
8.0K
The limit of detection (LOD) is the smallest amount of analyte that can be distinguished from the background noise. The LOD value corresponds to the concentration at which the analyte signal is three times larger than the standard deviation of the blank signal. Below this value, the analyte signal cannot be differentiated from the background noise. It is calculated by dividing the calibration slope by 3 times the standard deviation of the blank signals.
The LOD indicates the presence or absence...
The LOD indicates the presence or absence...
8.0K
Detection of Gross Error: The Q Test
6.9K
When one or more data points appear far from the rest of the data, there is a need to determine whether they are outliers and whether they should be eliminated from the data set to ensure an accurate representation of the measured value. In many cases, outliers arise from gross errors (or human errors) and do not accurately reflect the underlying phenomenon. In some cases, however, these apparent outliers reflect true phenomenological differences. In these cases, we can use statistical methods...
6.9K
What Are Outliers?
4.9K
Outliers are observed data points that are far from the least squares line. They have unusual values and need to be examined carefully. Though an outlier may result from erroneous data, at other times, it may hold valuable information about the population under study and should be included in the data. Hence, it is crucial to examine what causes a data point to be an outlier.
The z score is used to find outliers or unusual values. It should be noted that any values beyond -2 and +2 are...
The z score is used to find outliers or unusual values. It should be noted that any values beyond -2 and +2 are...
4.9K
