We present a quantitative performance analysis of a wide range of state-of-the-art object detection models, such as Mask R-CNN He et al.(2017)[8], RetinaNet Lin et al.(2017)[17] and EfficinetDet Tan et al.(2019)[28] in haze affected environments. This work uses two key performance metrics (Mean Average Precision and Localised Recall Precision) to provide a nuanced view of real world performance of these models in an on-road driving application. Our findings show that the presence of haze further exacerbates the performance differences between single-stage and multi-stage detection models. In addition, not all aspects of the model performance are affected equally. The inclusion of Local Recall Precision (LRP) Oksuz et al.(2018)[21] suggests that more recent models have much improved localisation performance even with similar false negative and false positive results. We also highlight some of the inherent limitations of Neural Network based approaches that could be addressed by Bayesian Neural Networks in the future.

我们对各种最先进的对象检测模型进行了定量性能分析，例如 Mask R-CNN He et al.(2017)[8]、RetinaNet Lin et al.(2017)[17]和 EfficinetDet Tan 等人（2019 年）[28] 在受雾霾影响的环境中。这项工作使用两个关键性能指标（平均平均精度和局部召回精度）来提供这些模型在道路驾驶应用程序中的真实世界性能的细微差别视图。我们的研究结果表明，雾霾的存在进一步加剧了单级和多级检测模型之间的性能差异。此外，并非模型性能的所有方面都受到同等影响。包含局部召回精度 (LRP) Oksuz 等人。(2018)[21] 表明，即使出现类似的假阴性和假阳性结果，更新的模型也大大提高了定位性能。我们还强调了基于神经网络的方法的一些固有局限性，未来贝叶斯神经网络可以解决这些局限性。