使用Tucker分解压缩视频的CNN内核：面向轻量级CNN应用
Compressing CNN Kernels for Videos Using Tucker Decompositions: Towards Lightweight CNN Applications
摘要：
卷积神经网络（CNN）是视觉计算领域的前沿技术。然而，CNN的一个主要问题是需要大量的浮点运算（FLOP）来执行大输入的卷积。当考虑将CNN应用于视频数据时，由于额外的时间维度，卷积滤波器变得更加复杂。这导致当相应的应用要被部署在移动设备（诸如智能电话、平板电脑、微控制器或类似物）上时的问题，指示较少的计算能力。Kim等人（2016）提出使用Tucker分解来压缩图像预训练网络的卷积核，以降低网络的复杂性，即FLOP的数量。在本文中，我们将上述方法推广到视频（和其他3D信号），并在THETIS数据集的修改版本上评估所提出的方法，该数据集包含个人进行网球拍摄的视频。我们表明，压缩的网络达到相当的精度，同时表明内存压缩的一个因素51。然而，实际的计算加速（因子1.4）并不符合我们理论上的预期（因子6）。
英文摘要：
Convolutional Neural Networks (CNN) are the state-of-the-art in the field of visual computing. However, a major problem with CNNs is the large number of floating point operations (FLOPs) required to perform convolutions for large inputs. When considering the application of CNNs to video data, convolutional filters become even more complex due to the extra temporal dimension. This leads to problems when respective applications are to be deployed on mobile devices, such as smart phones, tablets, micro-controllers or similar, indicating less computational power. Kim et al. (2016) proposed using a Tucker-decomposition to compress the convolutional kernel of a pre-trained network for images in order to reduce the complexity of the network, i.e. the number of FLOPs. In this paper, we generalize the aforementioned method for application to videos (and other 3D signals) and evaluate the proposed method on a modified version of the THETIS data set, which contains videos of individuals performing tennis shots. We show that the compressed network reaches comparable accuracy, while indicating a memory compression by a factor of 51. However, the actual computational speed-up (factor 1.4) does not meet our theoretically derived expectation (factor 6).
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。
