arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 1603.04871cs.CV
收藏

结合卷积层和递归层的最佳性能:一种用于语义分割的混合网络

Combining the Best of Convolutional Layers and Recurrent Layers: A Hybrid Network for Semantic Segmentation

Facebook(脸书);Google(谷歌);The University of Hong Kong(香港大学);University of Illinois at Urbana-Champaign(伊利诺伊大学厄巴纳-香槟分校);Carnegie Mellon University(卡内基梅隆大学)

Zhicheng Yan, Hao Zhang, Yangqing Jia, Thomas Breuel, Yizhou Yu

摘要:

语义分割的最新成果是由全卷积神经网络(FCN)建立的。FCS依赖于级联的卷积层和池化层来逐渐扩大神经元的感受野,从而以间接的方式对远程上下文依赖性进行建模。在这项工作中,我们提倡使用空间递归层(即ReNet层),它直接捕获全局上下文并导致改进的特征表示。我们通过构建一个朴素的深度ReNet(N-ReNet)来证明ReNet层的有效性,该ReNet在斯坦福背景数据集上实现了具有竞争力的性能。此外,我们将ReNet层与FCNs集成,并开发了一种新的混合深度ReNet(H-ReNet)。它具有一些显着的特性,包括全图像感受野,端到端训练和高效的网络执行。在PASCAL VOC 2012基准测试中,H-ReNet将最先进的分段、CRFasRNN和DeepParsing方法的结果分别提高了3.6%、2.3%和0.2%,并在20个对象类中的13个实现了最高的IoU。

英文摘要:

State-of-the-art results of semantic segmentation are established by Fully Convolutional neural Networks (FCNs). FCNs rely on cascaded convolutional and pooling layers to gradually enlarge the receptive fields of neurons, resulting in an indirect way of modeling the distant contextual dependence. In this work, we advocate the use of spatially recurrent layers (i.e. ReNet layers) which directly capture global contexts and lead to improved feature representations. We demonstrate the effectiveness of ReNet layers by building a Naive deep ReNet (N-ReNet), which achieves competitive performance on Stanford Background dataset. Furthermore, we integrate ReNet layers with FCNs, and develop a novel Hybrid deep ReNet (H-ReNet). It enjoys a few remarkable properties, including full-image receptive fields, end-to-end training, and efficient network execution. On the PASCAL VOC 2012 benchmark, the H-ReNet improves the results of state-of-the-art approaches Piecewise, CRFasRNN and DeepParsing by 3.6%, 2.3% and 0.2%, respectively, and achieves the highest IoUs for 13 out of the 20 object classes.

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。

↑