arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 1803.06508cs.CVcs.RO
收藏

MergeNet:一种用于小障碍发现的深网架构

MergeNet: A Deep Net Architecture for Small Obstacle Discovery

Microsoft, India(微软印度);International Institute of Information Technology Hyderabad (IIIT-H), India(印度海得拉巴国际信息技术学院(IIIT-H));Kohli Center of Intelligent Systems (KCIS), IIIT-H, India(印度海得拉巴国际信息技术学院智能系统科利中心(KCIS))

Krishnam Gupta, Syed Ashar Javed, Vineet Gandhi, K. Madhava Krishna

摘要:

我们在这里介绍一种名为MergeNet的新型网络架构,用于在自动驾驶的背景下发现道路场景中的小障碍物。该架构的基础在于以较少的数据量进行训练的核心考虑,因为小障碍物的物理设置和注释过程难以扩展。为了有效地利用有限的数据,我们提出了一个多阶段的训练过程,涉及权重共享,从RGBD输入和精炼阶段学习融合获得的互补特征的低和高级别特征的单独学习。该模型在Lost and Found数据集上进行了训练和评估,与之前基准测试使用的1000张图像相比,该模型仅使用135张图像就能够实现最先进的结果。此外,我们还将我们的结果与最近在6000张图像上训练的方法进行了比较,并表明我们的方法仅用1000个训练样本就实现了相当的性能。

英文摘要:

We present here, a novel network architecture called MergeNet for discovering small obstacles for on-road scenes in the context of autonomous driving. The basis of the architecture rests on the central consideration of training with less amount of data since the physical setup and the annotation process for small obstacles is hard to scale. For making effective use of the limited data, we propose a multi-stage training procedure involving weight-sharing, separate learning of low and high level features from the RGBD input and a refining stage which learns to fuse the obtained complementary features. The model is trained and evaluated on the Lost and Found dataset and is able to achieve state-of-art results with just 135 images in comparison to the 1000 images used by the previous benchmark. Additionally, we also compare our results with recent methods trained on 6000 images and show that our method achieves comparable performance with only 1000 training samples.

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。

↑