Background In this study, we propose the deep learning model-based framework to automatically delineate nasopharynx gross tumor volume (GTVnx) in MRI images. Methods MRI images from 200 patients were collected for training-validation and testing set. Three popular deep learning models (FCN, U-Net, Deeplabv3) are proposed to automatically delineate GTVnx. FCN was the first and simplest fully convolutional model. U-Net was proposed specifically for medical image segmentation. In Deeplabv3, the proposed Atrous Spatial Pyramid Pooling (ASPP) block, and fully connected Conditional Random Field(CRF) may improve the detection of the small scattered distributed tumor parts due to its different scale of spatial pyramid layers. The three models are compared under same fair criteria, except the learning rate set for the U-Net. Two widely applied evaluation standards, mIoU and mPA, are employed for the detection result evaluation. Results The extensive experiments show that the results of FCN and Deeplabv3 are promising as the benchmark of automatic nasopharyngeal cancer detection. Deeplabv3 performs best with the detection of mIoU 0.8529 ± 0.0017 and mPA 0.9103 ± 0.0039. FCN performs slightly worse in term of detection accuracy. However, both consume similar GPU memory and training time. U-Net performs obviously worst in both detection accuracy and memory consumption. Thus U-Net is not suggested for automatic GTVnx delineation. Conclusions The proposed framework for automatic target delineation of GTVnx in nasopharynx bring us the desirable and promising results, which could not only be labor-saving, but also make the contour evaluation more objective. This preliminary results provide us with clear directions for further study.

In this paper, a novel data Burst Assembly algorithm which based on control channel availability and traffic type is presented. In this mechanism, a burst is created only if its control packet can be transmitted and can be packed as different length according to traffic type to adapt the network requirement best dynamically. The variable Lstep which defined as the variation of the size threshold, can be chosen as the optimize value according to the traffic type and QoS requirement; by using the Lstep, this assembly algorithm exhibits a much better self-adaptive in the OBS Networks compared to the traditional burst adaptive algorithms, such as DTP and AthBA. The simulation results show that this algorithm can adaptively change the data burst size according to the offered load, and can help to reduce the packets loss rate and the ETE (End To End) delay of the bursts, thus significantly improve the network performance. Index Terms-adaptive, BA (Burst Assembly), burst loss rate, DTP (Data-length-Time-lag-Product), edge routerA. OBS Network Architecture B. Burst Assembly Algorithm

Correlated sources passing through broadcast channels is considered in this paper. Each receiver has access to correlated source side information and each source at the sender is kept secret from the unintended receiver. This communication model can be seen as generalizations of Tuncel's source over broadcast channel and Villard et al.'s source over wiretap channel. An outer bound for secure transmission region of arbitrarily correlated sources with the equivocation-rate levels is derived with ultra-low latency and used to prove capacity results for several classes of sources and channels.

No abstract