Search

Strong scattering media bring difficulties to imaging in many fields such as medicine and astronomy. The deep learning method has a powerful fitting ability, which can be better applied in reconstructing the target behind the scattering medium. But the detail of the reconstructed target is often inaccurate. In this paper, the skip connection is added in the neural network to improve the accuracy of the reconstructed detail. This network can combine pixel-level information with high-level semantic information, and the information missed during the encoding process can be more involved in the final reconstruction process. 1100 handwritten characters are used as the targets hidden behind the ground glass. It is found that the quality of the reconstructed target is different when the skip connection is added to the network at different scales. The feature map visualization method is used to help us analyze the role of the skip connection. Meanwhile, PSNR (Peak Signal to Noise Ratio) is also used as an objective evaluation standard to evaluate the quality of reconstructed targets. According to subjective and objective evaluation criteria, conclusions can be drawn that the detail of targets can be better retained when the skip connection is added between the convolutional layer corresponding to the feature map of the size 64*64, and the average PSNR can be enhanced 1 dB compared with the network without skip connections. This work provides a reference for the fusion methods of different scale features in the computational optical imaging.

The rapid growth of data volume brings big challenges to the data center computing, and energy efficiency is one of the most concerned problems. Researchers from various fields are now proposing solutions to green the data center operations. Power usage effectiveness metric plays an important role in the energy saving research. However, the exising usage effectiveness metrics focus on measuring the relationship between the total facility energy consumed and the IT equipment energy consumed, without reflecting the energy efficiency of applications. In this paper, we analyze the requirements of application-level metrics for power usage efficiency of the data centers, and propose two novel energy efficiency metrics to provide strong guidance and useful insight to data center design and optimization. We conduct comprehensive experiments in the practical data centers using BigDataBench, a big data benchmark suite, and the results demonstrate the rationality and efficiency of AxPUE in measuring the actual computation energy consumption in data centers., Comment: 8 pages, 4 figures, The First Workshop on Big Data Benchmarks, Performance Optimization, and Emerging hardware (BPOE 2013) In conjunction with 2013 IEEE International Conference on Big Data (IEEE Big Data 2013)

The performance analysis of manufacturing system originates from enterprise financial analysis, its purpose is to objectively analyze and evaluate the factors that can reflect the whole circumstance of the system by gathering all kinds of data in manufacturing system. This thesis mainly studies on the cash flow, which can reflect the managing and operating ability of modern manufacturing system, makes mapping between business indexes and manufacturing process indexes, builds evaluation model based on management accounting, which has the advantages of convenience, promptness and exactness. In the end, this thesis demonstrates the effectiveness of the model through example analysis.

Resource-based cities have to optimize their industry structures to fulfill sustainable development. During this procedure, they should supervise the developing efficiency of those dominant industries, it helps to judge whether the structure adjustment is on the right way. The beyond-efficiency DEA (BE-DEA)model is perfect to such evaluation, it can trace the dynamic efficiency, avoid the passiveness and waste of afterward evaluation. This paper applies this BE-DEA model to an oil resource city, evaluates its dominant industries' developing efficiency both horizontal and vertical. The result offers quantitative proof to adjust the target or policy, and demonstrates BE-DEA model's superiority.