Search

Recently, convolutional neural networks have attracted much attention due to its good performance in hyperspectral image classification. However, excessively increasing the depth of the network will lead to overfitting and vanishing gradient. Besides, previous networks rarely consider the related information among different convolution layers. In this paper we propose a hierarchical deep features adaptive fusion network (FAFNet) to address the above two problems. On the one hand, we use dense connectivity to overcome vanishing gradient and overfitting. On the other hand, we adaptively fuse different convolution layers by the learned weights which utilizing softmax to calculate. Experimental results on two well-known datasets demonstrate the excellent performance of the proposed method compared with other state-of-the-art methods.

Fine-grained images have similar global structure but exhibit variant local appearance. Bilinear pooling models have been proven to be effective in modeling different semantic parts and capturing the effective feature learning for fine-grained image classification. However, the bilinear models do not consider that convolutional neural networks (CNNs) may lose important semantic information during forward propagation, and feature interactions of different convolutional layers enhance feature learning which improves classification performance. Therefore, we propose a multi-layer weight-aware bilinear pooling method to model cross-layer object parts feature interaction as the feature representation, and different weights are assigned to each convolutional layer to adaptively adjust the outputs of the convolutional layers to highlight more discriminative features. The proposed method results in great performance improvement compared with previous state-of-the-art approaches. We demonstrate the effectiveness of our method on the CUB-200-2011 and FGVC-Aircraft datasets.

Using low-grade thermal energy instead of electricity to operate a refrigeration system can have important environmental benefits, especially when it is powered by a renewable energy source. Ejector refrigeration is one of the most promising technologies because of its relative simplicity and low capital cost when compared to an absorption refrigerator. An ejector heat pump is a heat-operated cycle capable of utilizing solar energy, waste energy, natural gas or hybrid sources (e.g. solar/gas). An ejector system basically consists of a generator, evaporator, condenser, ejector, expansion valve, and a pump. The ejector system has very few moving parts and so is simple in design. In addition, it has the potential of long life and, unlike vapour-compression systems, produces no noise or vibration. The system could be manufactured at relatively low cost, since inexpensive construction materials may be used. Although they have a relatively low coefficient of performance compared to air-conditioning systems using mechanical compressors, the ejector cooling technologies have attracted extensive attentions with ever-increasing awareness and pressures for protecting the environment and have achieved significant improvement in coefficient of performance as compared to other systems. The continuous developments in solar collector technology open the way to the effective utilization of solar energy to power the ejector systems and utilization of environmental friendly refrigerants is also the major concern. This chapter introduces the principle of the ejector, basic ejector cycle, solar-driven ejector system and its operating. The refrigerants, solar collectors, and PCM heat storage for solar ejector system applications are also introduced. A complete solar ejector air-conditioning system used in a building is presented in this chapter.