Academic literature on the topic 'L2 Normalization'

Transmission, as a critical part of vehicles, is the hub of power transmission and the core of controlling speed change. Condition monitoring and diagnosis of transmissions have become an effective tool to ensure vehicle safety travelling. The intelligent fault diagnosis strategy using artificial intelligent methods has been studied and applied for gearbox fault diagnosis. However, most algorithms cannot guarantee both accuracy and training efficiency. In this paper, fast convolutional sparse filtering based on convolutional activation and feature normalization is proposed for gearbox fault diagnosis without any time-consuming preprocessing. In fast convolutional sparse filtering, the features of samples are optimized instead of local features, which could obviously reduce the dimension and construction time of the Hessian matrix. In addition, the output features are equally active to guarantee that all features have similar contributions. The l2-norm of the training features is recorded and used for pseudo-normalization of the test features. The proposed fast convolutional sparse filtering is validated by a bearing fault dataset and a planetary gear fault dataset. Verification results confirm that fast convolutional sparse filtering is a promising tool for fault diagnosis, which has obviously improved the diagnosis accuracy, training efficiency, and robustness and provides the greater advantage of handling large-scale datasets.

This paper provides an efficient way of addressing the problem of detecting or estimating the 6-Dimensional (6D) pose of objects from an RGB image. A quaternion is used to define an object′s three-dimensional pose, but the pose represented by q and the pose represented by -q are equivalent, and the L2 loss between them is very large. Therefore, we define a new quaternion pose loss function to solve this problem. Based on this, we designed a new convolutional neural network named Q-Net to estimate an object’s pose. Considering that the quaternion′s output is a unit vector, a normalization layer is added in Q-Net to hold the output of pose on a four-dimensional unit sphere. We propose a new algorithm, called the Bounding Box Equation, to obtain 3D translation quickly and effectively from 2D bounding boxes. The algorithm uses an entirely new way of assessing the 3D rotation (R) and 3D translation rotation (t) in only one RGB image. This method can upgrade any traditional 2D-box prediction algorithm to a 3D prediction model. We evaluated our model using the LineMod dataset, and experiments have shown that our methodology is more acceptable and efficient in terms of L2 loss and computational time.

Background and aims Fecal microbial transplantation (FMT) is empirically implemented in horses with colitis to facilitate resolution of diarrhea. The purpose of this study was to assess FMT as a clinical treatment and modulator of fecal microbiota in hospitalized horses with colitis. Methods A total of 22 horses with moderate to severe diarrhea, consistent with a diagnosis of colitis, were enrolled at two referral hospitals (L1: n = 12; L2: n = 10). FMT was performed in all 12 patients on 3 consecutive days at L1, while treatment at L2 consisted of standard care without FMT. Manure was collected once daily for 4 days from the rectum in all colitis horses, prior to FMT for horses at L1, and from each manure sample used for FMT. Fecal samples from 10 clinically healthy control horses housed at L2, and 30 healthy horses located at 5 barns in regional proximity to L1 were also obtained to characterize the regional healthy equine microbiome. All fecal microbiota were analyzed using 16S amplicon sequencing. Results and conclusions As expected, healthy horses at both locations showed a greater α-diversity and lower β-diversity compared to horses with colitis. The fecal microbiome of healthy horses clustered by location, with L1 horses showing a higher prevalence of Kiritimatiellaeota. Improved manure consistency (lower diarrhea score) was associated with a greater α-diversity in horses with colitis at both locations (L1: r = -0.385, P = 0.006; L2: r = -0.479, P = 0.002). Fecal transplant recipients demonstrated a greater overall reduction in diarrhea score (median: 4±3 grades), compared to untreated horses (median: 1.5±3 grades, P = 0.021), with a higher incidence in day-over-day improvement in diarrhea (22/36 (61%) vs. 10/28 (36%) instances, P = 0.011). When comparing microbiota of diseased horses at study conclusion to that of healthy controls, FMT-treated horses showed a lower mean UniFrac distance (0.53±0.27) than untreated horses (0.62±0.26, P<0.001), indicating greater normalization of the microbiome in FMT-treated patients.

This paper describes the detection of coconut trees using very-high-resolution optical satellite imagery. The satellite imagery used in this study was a panchromatic band of Pleiades imagery with a spatial resolution of 0.5 metres. The authors proposed the use of a histogram of oriented gradients (HOG) algorithm as the feature extractor and a support vector machine (SVM) as the classifier for this detection. The main objective of this study is to find out the parameter combination for the HOG algorithm that could provide the best performance for coconut-tree detection. The study shows that the best parameter combination for the HOG algorithm is a configuration of 3 x 3 blocks, 9 orientation bins, and L2-norm block normalization. These parameters provide overall accuracy, precision and recall of approximately 80%, 73% and 87%, respectively.