Search

This paper proposes a real-time fire detection framework DeepFireNet that combines fire features and convolutional neural networks, which can be used to detect real-time video collected by monitoring equipment. DeepFireNet takes surveillance device video stream as input. To begin with, based on the static and dynamic characteristics of fire, a large number of non-fire images in the video stream are filtered. In the process, for the fire images in the video stream, the suspected fire area in the image is extracted. Eliminate the influence of light sources, candles and other interference sources to reduce the interference of complex environments on fire detection. Then, the algorithm encodes the extracted region and inputs it into DeepFireNet convolution network, which extracts the depth feature of the image and finally judges whether there is a fire in the image. DeepFireNet network replaces 5×5 convolution kernels in the inception layer with two 3×3 convolution kernels, and only uses three improved inception layers as the core architecture of the network, which effectively reduces the network parameters and significantly reduces the amount of computation. The experimental results show that this method can be applied to many different indoor and outdoor scenes. Besides, the algorithm effectively meets the requirements for the accuracy and real-time of the detection algorithm in the process of real-time video detection. This method has good practicability.

Portable box volume measurement has always been a popular issue in the intelligent logistic industry. This work presents a portable system for box volume measurement that is based on line-structured light vision and deep learning. This system consists of a novel 2 &times, 2 laser line grid projector, a sensor, and software modules, with which only two laser-modulated images of boxes are required for volume measurement. For laser-modulated images, a novel end-to-end deep learning model is proposed by using an improved holistically nested edge detection network to extract edges. Furthermore, an automatic one-step calibration method for the line-structured light projector is designed for fast calibration. The experimental results show that the measuring range of our proposed system is 100&ndash, 1800 mm, with errors less than &plusmn, 5.0 mm. Theoretical analysis indicates that within the measuring range of the system, the measurement uncertainty of the measuring device is &plusmn, 0.52 mm to &plusmn, 4.0 mm, which is consistent with the experimental results. The device size is 140 mm &times, 35 mm &times, 35 mm and the weight is 110 g, thus the system is suitable for portable automatic box volume measurement.

In the collaborative evolution of biomedical ontology, participants are a sizeable, balanced mix of scholars and physicians, all of whom are experienced in the biomedical domain and who represent diverse viewpoints, experiences, and backgrounds. In the field of collaborative evolution of biomedical ontology on large-scale ontology, there exists inevitable conflicts, which may cause the inconsistent ontology. In this paper, a new method to detect conflicts in ontology evolution is presented, which classifies conflicts as three groups: internal inconsistencies conflicts in change sequence, direct conflicts between the sequences and Inconsistent conflict between the sequences. For different conflict, high effective detecting algorithms are presented with evaluation. Before the conflicts detecting, semantic extended rules are employed to depict the evolution requirements of the participants. In particular, we discuss the situation where maximum consistent changing subsequence is needed if there are inconsistent conflicts between changing subsequences. We also show how detecting algorithms could be taken in the collaborative evolution of medicine ontology. As a result, internal and mutual inconsistencies can be detected from change sequences. And if there are conflicts between the sequences, the algorithm will provide the maximum consistent changing subsequence as the evolution basis. The designed experiments verify our approach and achieve the expected results.

In this paper, a family of high efficiency high step-up dc-dc converter is presented, which inherits all advantages of both switched-inductor dc-dc converter and coupled-inductor dc-dc converter. The proposed converters can achieve high step-up capability with the proper duty cycle D (high or small D leads to the bad efficiency), obtain soft-switching conditions of active devices (including switch and rectifier diodes), and recycle the energy of leakage inductance to the output capacitor, which results in the high efficiency. The switch of converter operates in ZCS turn-on condition, which decreases the switching loss and improves the switch operating environment. The rectifier diodes cut off with ZCS condition, which improves the reverse recovery characteristics of diodes. The voltage relationships, operating principle, and operating modes are described in detail. Finally, experimental results verity the validity of the proposed switched-coupled inductor boost converter.

The goal of the Gene Ontology is to provide a controlled vocabulary that can be applied to all organisms, even as knowledge of gene and protein roles in cells is accumulating and changing. Further, in the process of co-evolution, conflicts are unavoidable among ontology changing sequences. We use the N ⨉ M matrix to model the changing sequences in Gene ontology evolution which lead to a result ontology. According the users’ estimate of the changing sequences, we propose the updating algorithm to update the matrix as mentioned earlier. The new matrix is as the input of the MUOUE((A,R) algorithm to calculate the minimal hitting sequence set of high-potential candidates that were subject to a subsequent Bayesian reasoning approach to rank them. The instances show that our approach can be used to diagnosis the conflicts, especially under the opening environment. The diagnosis approach we proposed in this paper can provide support for the co-evolution of Gene Ontology.

In order to further improve fractional-pel interpolation image quality of video sequence with different resolutions and reduce algorithm complexity, the fractional-pel interpolation algorithm based on adaptive filter (AF_FIA) is proposed. This algorithm adaptively selects the interpolation filters with different orders according to the three video sequence regions with different resolutions; in the three video sequence regions with different resolutions, the high-order interpolation filter is replaced by low-order interpolation filter according to the correlation between pixels to realize the adaptive selection of filter. The complexity analysis results show that compared with other algorithms, this algorithm reduces space complexity and computation complexity, thus reducing the storage access and coding time. The simulation results indicate that compared with other algorithms, this algorithm has good coding performance and robustness for video sequences with different resolutions.

As the core link and an important content, credit risk assessment plays vital special role for moderate operation of commercial banks. It is important to work out credit risk assessment model of commercial banks. A system of credit risk assessment which contains five aspects including 25 indicators is proposed. Based on establishment of credit risk assessment model, empirical analysis is carried out. The result shows this model has high accuracy of predication, better reliability and strong assessment ability.