Search

This paper gave an improved ensemble model for Chinese text classification based on CNN and RNN structure. We first changed the bidirectional LSTM structure in the TextRCNN model to a bidirectional GRU structure and added a learning rate reduction mechanism into it. Then TextCNN model was also added into a learning rate reduction mechanism. Finally, our ensemble model was constructed through equal weight voting between the improved TextRCNN and improved TextCNN models. Experimental results showed that compared with TextCNN [7], TextRCNN [8], LSTM+attention [15], DPCNN [5] and FastText [6], the classification error rate of our model on SINA_P News dataset dropped by 8.36%, 13.2%, 20.23%, 7.33% and 29.17% respectively; dropped by 8.28%, 3.98%, 10.68% and 5.86% and rose by 3.8% on THUC_P news dataset. Hence, our ensemble model has higher classification accuracy than most other text classification models and is a competitive method.

In order to reduce the cost of electrode active material, a three-dimensional (3D) sponge-like Co/Co(OH) 2 nanocomposite was synthesized by a simple NaBH 4 reduction method at room temperature. For supercapacitor application, the Co/Co(OH) 2 nanocomposite exhibits an excellent specific capacity of 1048 F g −1 at a current density of 1 A g −1 in a potential range of 0–0.45 V from the charge/discharge test, and the retention of capacitance is 70.6% after 5000 cycles. The sponge-like Co/Co(OH) 2 nanocomposite displays high supercapacitor performance, excellent rate capability and cycling stability.

La correcta medicion de los parametros de las plumas es una cuestion clave para garantizar la calidad de la fabricacion de volantes de badminton. Sin embargo, la deteccion manual sigue siendo el principal metodo de medicion de parametros utilizado en la actualidad. Por eso, este estudio propone un principio de diseno y un metodo de medicion, mediante una estructura basica con una sola camara multi-sistema para la medicion de parametros. En primer lugar, sobre la base del principio de imagen en lente convexa, con el dispositivo de medicion propuesto, se tomaron las imagenes de las plumas dentro de un espejo a diferentes angulos. Se obtuvieron tambien las formulas para medir su curvatura, caida, ancho y espesor. En segundo lugar, se analizo la relacion entre la caida y la profundidad de campo. El analisis indico que todas la plumas con diferentes niveles de caida debian situarse dentro de los limites de la profundidad de campo para asegurar que la camara obtenia imagenes claras. Por ultimo, los valores medidos se compararon con los resultados obtenidos por medicion manual con un calibre normal. La curvatura, caida, espesor y ancho de los trozos de pluma pueden medirse con el dispositivo propuesto de una sola camara segun los resultados practicos obtenidos. Los resultados de este estudio pueden servir como referencia para aumentar el rendimiento y la calidad de la pluma de badminton y reducir el coste del dispositivo y los errores humanos. Keywords: imagen digital de pluma, espejo plano, curvatura y caida, medicion de parametros

To improve the safety, accuracy and effectiveness of surgery, while also reducing surgical trauma, three-dimensional spatial positioning of the surgical instruments can be achieved through the accuracy of integration of medical imaging to guide their operation during surgery. Therefore, prior to surgery, registration is required to establish the relationship between the actual positioning of the surgical instruments and their imaged position. The accuracy of such registration directly determines the precision of navigation. Taking into account the shortcomings and deficiencies of the current fiducial marker registration methods, this paper proposes an automatic fiducial marker extraction and matching algorithm. In order to facilitate automatic recognition and extraction of the fiducial markers the authors of this paper designed fiducial marker with CT values and shapes that provide recognizable characteristics which help to meet the demands of the positioning requirements. The extraction of such fiducial markers can be easily automated, significantly reducing the difficulty in developing an extraction algorithm. After fiducial marker extraction the registration algorithm and the Iterative Closest Point (ICP) algorithm can be combined to achieve matching between the fiducial marker and point set model of the fiducial marker. According to the matching results indicating the positional relationship between the each fiducial marker and its retroreflective sphere, the center coordinate of the retroreflective sphere in the image can be acquired, hence providing an easy way to undertake registration of the different spatial representations of the patient.

Patients who has fecal incontinence suffer physical and mental painfulness because they can neither actively control any forms of defecation nor detect the state of intestinal contents. It is significantly meaningful to study artificial anus sphincter with “artificial sensory nerves”. In this paper, reflective infrared (IR) sensor, with different responses to the solutes, can simulate the sensory nerves to detect intestinal contents to distinguish different solutes according to the voltage values of the different reflective responses to the IR of the solutes. The study meets the requirements as expected because the outcome shows that the voltage signals of different solution differ dramatically which can detect whether there are any intestinal contents and distinguish the states of them. It turns out to be an effective method that the reflective IR sensor can simulate the sensory nerves to sense intestinal contents.