the need to acquire high performance deep neural network models is a research trend in recent years. Many examples have shown that achieving high validation accuracies require a very large number of parameters in most...
ISBN:
(数字)9781728126104
ISBN:
(纸本)9781728126111
the need to acquire high performance deep neural network models is a research trend in recent years. Many examples have shown that achieving high validation accuracies require a very large number of parameters in most cases and therefore, the space used to store these models becomes very large. this may be a disadvantage on small storage size and low performance CPU edge devices during image processing that are embedded with neural networks for object recognition tasks. In this paper, we investigate the effect of input images which are partially compressed using the Discrete Cosine Transform (DCT) algorithm on two different Convolutional Neural Network (CNN) performances, known as CNN-C (large model) and CNN-RC3 (small model). DCT is used to reduce some data redundancies but also the risk of losing valuable features for the network to learn efficiently. However, the results show that both CNN architectures with DCT features perform as well as with raw image data, concluding that a properly designed CNN model can still achieve high performance on further compressed images regardless of its information reductions.
Peer-to-Peer (P2P) communications have become a popular alternative solution to provide large-scale video-on-demand (VoD) services. Recent approaches are designed for streaming applications in a unicast infrastructure...
详细信息
Peer-to-Peer (P2P) communications have become a popular alternative solution to provide large-scale video-on-demand (VoD) services. Recent approaches are designed for streaming applications in a unicast infrastructure. As the successful deployment of IP broadcast delivery, the system could have a further improvement in terms of cost effective and scalability when broadcasting scheme can be coupled with P2P paradigm in a suitable manner. In this paper, we develop a possible solution for building a VoD system using existing broadcasting protocol coupled with P2P paradigm denoted Broadcast-based Peer-to-Peer Protocol with Erasure Correcting Codes (BPP-ECC). A video is partitioned and distributed to a number of peer servers for broadcasting. Because of the dynamic nature of P2P application, a fault tolerant mechanism based on erasure correcting codes is employed to increase the system reliability. To avoid the disruption of service, a central server is still deployed in the system to take over the service when the system does not have sufficient peer server to sustain the service. We thus consider the trade-off between the number of peer servers required and the bandwidth requirement of the central server. An analytical model is developed to determine the minimum number of peer servers required for the system. the results showed that 51 peers, each of which has the availability of 0.6 and has a storage of 288 seconds, are enough to leverage the workload of the central server up to 95%, when the startup delay of the system is 5 minutes.
暂无评论