Smoke is an informative indicator of early fire and gas leakage. Segmenting the smoke from images can provide detailed information about the smoke volume, dispersion direction, and source location, which has significant implications considering the proliferation of video surveillance systems in cities. Focusing on smoke segmentation in the urban scene, we designed a dual-branch segmentation model, named SmokeSeger, which couples a transformer branch and a convolutional neural network (CNN) branch to enhance the representation of both global and local features. To address the lack of real-scene smoke datasets, we built an urban scene smoke segmentation dataset containing 3217 images of fire smoke and exhaust emissions with accurate annotations. Experiments validate that the SmokeSeger outperforms other mainstream segmentation methods on the proposed dataset. Visualization of attention maps reveals that the model could effectively capture the semantic relationship between the smoke and the corresponding source, which benefits the discrimination between smoke and smoke-like objects. More details available athttps://github.com/VisAcademic/SmokeSeger.

Smoke detection has been a research focus due to its application value in fire and toxic gas leakage alarms. Here we formulate a novel research paradigm for in-depth smoke analysis: modeling the smoke plume in an image with two semantic keypoints, i.e., the start-point (where the smoke comes from) and the end-point (where the smoke spreads), and localizing the keypoints through a heatmap-based detection method. A specialized dataset is developed for smoke keypoint detection, collecting images online and manually annotating the keypoints. Based on the dataset, we propose a Transformer-based model called SmokePose that employs a hierarchical Transformer encoder and a pure Transformer decoder to detect smoke keypoints in an end-to-end manner. We demonstrated the performance of the proposed SmokePose with comparative experiments and ablation studies. A further discussion on the visualization of the attention maps helps to understand the mechanism of SmokePose and to reveal essential image clues for smoke keypoint detection.

Touch gesture recognition (TGR) plays a pivotal role in many applications, such as socially assistive robots and embodied telecommunication. However, one obstacle to practicality of existing TGR methods is the individual disparities across subjects. Moreover, a deep neural network trained with multiple existing subjects can easily lead to overfitting for a new subject. Hence, how to mitigate the discrepancies between the new and existing subjects and establish a generalized network for TGR is a significant task to realize reliable human–robot tactile interaction. In this article, a novel framework for Multisource domain Adaptation via Shared-Specific feature projection (MASS) is proposed, which incorporates intradomain discriminant, multidomain discriminant, and cross-domain consistency into a deep learning network for cross-subject TGR. Specifically, the MASS method first extracts the shared features in the common feature space of training subjects, with which a domain-general classifier is built. Then, the specific features of each pair of training and testing subjects are mapped and aligned in their common feature space, and multiple domain-specific classifiers are trained with the specific features. Finally, the domain-general classifier and domain-specific classifiers are ensembled to predict the label for the touch samples of a new subject. Experimental results performed on two datasets show that our proposed MASS method achieves remarkable results for cross-subject TGR. The code of MASS is available athttps://github.com/AI-touch/MASS.

Olfactory-induced electroencephalogram (EEG) signal classification is of great significance in a variety of fields, such as disorder treatment, neuroscience research, multimedia applications and brain–computer interface. In this paper, a trapezoid difference-based electrode sequence hashing method is proposed for olfactory EEG signal classification. First, an [Formula: see text]-layer trapezoid feature set whose size ratio of the top, bottom and height is 1:2:1 is constructed for each frequency band of each EEG sample. This construction is based on [Formula: see text] optimized power-spectral-density features extracted from [Formula: see text] real electrodes and [Formula: see text] nonreal electrode’s features. Subsequently, the [Formula: see text] real electrodes’ sequence (ES) codes of each layer of the constructed trapezoid feature set are obtained by arranging the feature values in ascending order. Finally, the nearest neighbor classification is used to find a class whose ES codes are the most similar to those of the testing sample. Thirteen-class olfactory EEG signals collected from 11 subjects are used to compare the classification performance of the proposed method with six traditional classification methods. The comparison shows that the proposed method gives average accuracy of 94.3%, Cohen’s kappa value of 0.94, precision of 95.0%, and F1-measure of 94.6%, which are higher than those of the existing methods.

For the classifiaction tasks within two classes, a feature extraction method combining the principal component analysis (PCA) and the linear discriminant analysis (LDA) is adopted, and an improved extreme learning machine (ELM), i.e., the near extreme learning machine (NELM classification algorithm), is presented. To further improve the classification performance, a voting-NELM (VNELM) is proposed. To examine the performance of our proposed classification algorithm, two different tests were carried out: slow cortical potential (SCP) signal classification and Chinese liquor (true or false) recognition. Experimental results reveal that for the SCP signal classification using the BCI competition II dataset Ia, an accuracy of 93.52% is obtained through the VNELM algorithm, better than that (92.30%) of the state-of-the-art method (i.e., the best improved ELM algorithm, V-ELM). When applied the VNELM algorithm to the Chinese liquor recognition, all the single-sensor-based classification results are better than that of the ELM and the V-ELM, and a better average accuracy of 99.25% is obtained based on the multi-sensor response signals, increasing the accuracy by 26% and 6.25% from that (73.25% and 93.00%) of the ELM and the V-ELM, respectively.

Classification of slow cortical potential (SCP) signals is crucial for brain-computer interface (BCI) systems. This letter presents a new scheme to improve the classification performance of SCP signals. It consists of two parts: first, by fitting the wavelet coefficients of SCP signals with a second-order polynomial, the SCP trends are extracted; and second, a voting system based on the optimal training parameters of the support vector machines is developed to enhance the classification accuracy (CA). Experimental results reveal that the proposed scheme outperforms the state-of-the-art methods. The CA improvements for the dataset Ia of the BCI competition II and the TJU dataset (the dataset was collected in Tianjin University, termed TJU dataset) are reported.