Real-Time Semantic Segmentation with Dual Encoder and Self-Attention Mechanism for Autonomous Driving.

As the techniques of autonomous driving become increasingly valued and universal, real-time semantic segmentation has become very popular and challenging in the field of deep learning and computer vision in recent years. However, in order to apply the deep learning model to edge devices accompanying sensors on vehicles, we need to design a structure that has the best trade-off between accuracy and inference time. In previous works, several methods sacrificed accuracy to obtain a faster inference time, while others aimed to find the best accuracy under the condition of real time. Nevertheless, the accuracies of previous real-time semantic segmentation methods still have a large gap compared to general semantic segmentation methods. As a result, we propose a network architecture based on a dual encoder and a self-attention mechanism. Compared with preceding works, we achieved a 78.6% mIoU with a speed of 39.4 FPS with a 1024 × 2048 resolution on a Cityscapes test submission.

Schistosoma egg antigens suppress LPS-induced inflammation in human IMR-90 cells by modulation of JAK/STAT1 signaling.

BACKGROUND: The regulation of the balance between inflammatory and anti-inflammatory events during the treatment of pulmonary infection is very important. Soluble Schistosoma egg antigens (SEA) can effectively inhibit the expression of cytokines during hepatic acute inflammation. However, the mechanisms by which these proteins suppress the inflammatory responses in lung cells remain unclear. The purpose of this study was to investigate the ability of SEA to inhibit pulmonary inflammation. METHODS: The effects of SEA were investigated in LPS-treated lung IMR-90 cells. The involvement of the JAK/STAT-1 signaling pathway in these effects was evaluated by employing CBA assays, quantitative polymerase chain reaction, and western blotting experiments. RESULTS: Pretreatment of IMR-90 cells with appropriate concentrations of SEA protected cells against the cytotoxic effects of LPS-induced inflammation in a time-dependent manner. SEA pretreatment significantly attenuated the LPS-induced activation of the JAK/STAT1 signaling pathway, including the upregulation of JAK1/2 and STAT1, as well as the production of inflammatory cytokines. The level of phosphorylated STAT1 gradually declined in response to increasing concentrations of SEA. Based on these findings, we hypothesize that SEA-induced anti-inflammatory effects initiate with the downregulation of the IFN-γ-JAK-STAT1 signaling pathway, resulting in the attenuation of LPS-induced inflammation in IMR-90 cells. CONCLUSION: Our study is the first to demonstrate the anti-inflammatory activity of SEA in an in vitro model of pulmonary inflammation, involving the modulation of JAK/STAT1 signaling. We propose SEA as potential therapeutic or preventive agents for the selective suppression of STAT1 and the control of inflammatory response in lung IMR-90 cells.

Pure FPGA Implementation of an HOG Based Real-Time Pedestrian Detection System.

In this study, we propose a real-time pedestrian detection system using a FPGA with a digital image sensor. Comparing with some prior works, the proposed implementation realizes both the histogram of oriented gradients (HOG) and the trained support vector machine (SVM) classification on a FPGA. Moreover, the implementation does not use any external memory or processors to assist the implementation. Although the implementation implements both the HOG algorithm and the SVM classification in hardware without using any external memory modules and processors, the proposed implementation's resource utilization of the FPGA is lower than most of the prior art. The main reasons resulting in the lower resource usage are: (1) simplification in the Getting Bin sub-module; (2) distributed writing and two shift registers in the Cell Histogram Generation sub-module; (3) reuse of each sum of the cell histogram in the Block Histogram Normalization sub-module; and (4) regarding a window of the SVM classification as 105 blocks of the SVM classification. Moreover, compared to Dalal and Triggs's pure software HOG implementation, the proposed implementation's average detection rate is just about 4.05% less, but can achieve a much higher frame rate.

Emotion identification using extremely low frequency components of speech feature contours.

The investigations of emotional speech identification can be divided into two main parts, features and classifiers. In this paper, how to extract an effective speech feature set for the emotional speech identification is addressed. In our speech feature set, we use not only statistical analysis of frame-based acoustical features, but also the approximated speech feature contours, which are obtained by extracting extremely low frequency components to speech feature contours. Furthermore, principal component analysis (PCA) is applied to the approximated speech feature contours so that an efficient representation of approximated contours can be derived. The proposed speech feature set is fed into support vector machines (SVMs) to perform multiclass emotion identification. The experimental results demonstrate the performance of the proposed system with 82.26% identification rate.

Music emotion detection using hierarchical sparse kernel machines.

For music emotion detection, this paper presents a music emotion verification system based on hierarchical sparse kernel machines. With the proposed system, we intend to verify if a music clip possesses happiness emotion or not. There are two levels in the hierarchical sparse kernel machines. In the first level, a set of acoustical features are extracted, and principle component analysis (PCA) is implemented to reduce the dimension. The acoustical features are utilized to generate the first-level decision vector, which is a vector with each element being a significant value of an emotion. The significant values of eight main emotional classes are utilized in this paper. To calculate the significant value of an emotion, we construct its 2-class SVM with calm emotion as the global (non-target) side of the SVM. The probability distributions of the adopted acoustical features are calculated and the probability product kernel is applied in the first-level SVMs to obtain first-level decision vector feature. In the second level of the hierarchical system, we merely construct a 2-class relevance vector machine (RVM) with happiness as the target side and other emotions as the background side of the RVM. The first-level decision vector is used as the feature with conventional radial basis function kernel. The happiness verification threshold is built on the probability value. In the experimental results, the detection error tradeoff (DET) curve shows that the proposed system has a good performance on verifying if a music clip reveals happiness emotion.