arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 1908.03621cs.CV
收藏

一种用于概率目标检测的Mask-RCNN基线

A Mask-RCNN Baseline for Probabilistic Object Detection

UNC-Chapel Hill(北卡罗来纳大学教堂山分校)

Phil Ammirato, Alexander C. Berg

摘要:

概率目标检测挑战赛在新的合成图像数据集上使用新的评估指标--基于概率的检测质量(PDQ)来评估目标检测方法。我们提出了我们的挑战,一个微调版本的Mask-RCNN与一些额外的后处理。我们的方法以用户名pammirato提交,目前在排行榜上排名第二,得分为21.432,同时还实现了最高的空间质量和平均整体检测质量。我们希望这种方法可以提供一些洞察如何检测器设计的平均精度(mAP)评估PDQ下的行为,以及为未来的工作强有力的基线。

英文摘要:

The Probabilistic Object Detection Challenge evaluates object detection methods using a new evaluation measure, Probability-based Detection Quality (PDQ), on a new synthetic image dataset. We present our submission to the challenge, a fine-tuned version of Mask-RCNN with some additional post-processing. Our method, submitted under username pammirato, is currently second on the leaderboard with a score of 21.432, while also achieving the highest spatial quality and average overall quality of detections. We hope this method can provide some insight into how detectors designed for mean average precision (mAP) evaluation behave under PDQ, as well as a strong baseline for future work.

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。

↑