Abstract. In order to measure the weight of yak more conveniently and effectively, based on BP neural network, this paper provides a portable dynamic weighing System for yak. The wireless transmission mode is adopted between the acquisition module and the instrument in this system and weighing platform is added a handle and a small roller table, which overcome the shortcomings of traditional weighing station that cannot move. In order to further improve the problem of traditional weighing station's low accuracy, using the smoothing mean filter to denoise the original data and according to the output of the shear beam type weighing sensor and the speed of yak, the BP neural model is established, thus the static weight of yak being obtained. By many experiments in matlab, the results show that this system achieves the measurement accuracy of dynamic weighing system and ensures that it can be achieved technically, which has good practical value.

Recent years, deep learning in the image identification has made great progress, showing good application prospects in medical image reading. Diabetic Retinopathy (DR) is an eye disease due to diabetes, which is the most ordinary cause of blindness. Traditional diabetic retinopathy detection is a manual and time-consuming and labor-intensive process, which requires a highly experienced clinician to examine and evaluate the digital color fundus photos of the retina. Therefore, it is crucial to use the deep learning technique to achieve automatic detection of diabetic retinopathy. In this paper, we proposed a diabetic retinopathy detection method based on deep learning and proposed a network structure named multi-self-attention. At first, the image features were extracted through the InceptionV3 model, and then the feature maps was directly generated. Secondly, the feature maps, which can reflect condition of retina, will be input into multi-self-attention network structure, to calculate multi-self-attention feature. Finally, by convolutional layer and fully connected layer, the stage results about diabetic retinopathy will be obtained. With the experiments in TensorFlow framework , the effectiveness of multi-self-attention network structure for feature extraction and classification is proved.

Convolutional Neural Networks (CNNs) have performed very well on image classification tasks, but CNNs is insensitive to detailed image information and requires a large amount of training data and time. Capsule Networks(CapsNets) can solve this problem very well, but the Baseline CapsNet model is very shallow, and the extraction of low-level features is not enough. We propose a Multi-Scale Capsule Network (Multi-Scale CapsNet), by extracting the low-level features of images with multi-channel convolution of multiple convolution kernels, so extracted features are more diverse, then passing from the bottom layer to the upper layer in the form of a "capsule", which encapsulats the multidimensional features of the image in the form of a vector, thus the features are saved in the network, rather than being recovered after being lost. In the German Traffic Sign Recognition Benchmark(GTSRB), we obtained competitive results with the accuracy of 99.4%, which is better than the human performance of 98.81% and the Multi-Scale Convolutional Neural Network(MS-CNN) of 97.33%.

Abstract. To get self-adaptive thresholds, the edge detection algorithm of Canny operator is improved in OpenCV library under Linux platform in this paper. Firstly, the gradient graph of the gray image and the maximum value of the gradient are obtained. Then the histogram is calculated and the pixel points corresponding to the gradient values are obtained via the traversion of histogram. Finally, the high and low thresholds are calculated to determine edge. Consequently, there comes the new self-adaptive thresholds algorithm of Canny operator. In the meantime, by simulating on the robotic fish in the water, the new algorithm is compared with the traditional Canny operator in the effect of the output image and the Peak Signal to Noise Ratio. What's more, the same picture is tested to compare time in the Matlab platform and OpenCV platform. Such comparisons show that the new algorithm has more flexibility and efficiency.

Abstract:The edge is one of most significant features of the images, which is the basis of image analysis and recognition. When it comes to the segmentation, measurement of the object, the edge extraction and the noise resistance are of particular importance. In order to achieve the effective extraction of the image edge, this paper presents a new algorithm for the detection of image edge width. This method is based on the principle that the blurred image will lead to edge diffusion and is divided into two processing stages. The first step is aimed at calculating the maximum value of local gradients of the image, obtaining the rough edge information map of the image, and then obtaining the finer edge information map via the adaptive dual threshold selection method. In the second stage, the probability curve on the edge width of the graph is plotted, the skewness and variance are calculated, and then we can obtain the decline factor of the curve. Finally, we define the decline factor as the adjustment coefficient which is multiplied by the average width of the image edge and obtain the articulation value. Having analyzed the of sharpness of the image edge width which is extracted in both objective and subjective ways, the experimental results show that the edge width information extracted by this algorithm is close to the subjective evaluation of human eyes, further illustrating the feasibility of the proposed algorithm in this paper.