KepSalinst: Using Peripheral Points to Delineate Salient Instances.

Salient instance segmentation (SIS) is an emerging field that evolves from salient object detection (SOD), aiming at identifying individual salient instances using segmentation maps. Inspired by the success of dynamic convolutions in segmentation tasks, this article introduces a keypoints-based SIS network (KepSalinst). It employs multiple keypoints, that is, the center and several peripheral points of an instance, as effective geometrical guidance for dynamic convolutions. The features at peripheral points can help roughly delineate the spatial extent of the instance and complement the information inside the central features. To fully exploit the complementary components within these features, we design a differentiated patterns fusion (DPF) module. This ensures that the resulting dynamic convolutional filters formed by these features are sufficiently comprehensive for precise segmentation. Furthermore, we introduce a high-level semantic guided saliency (HSGS) module. This module enhances the perception of saliency by predicting a map for the input image to estimate a saliency score for each segmented instance. On four SIS datasets (ILSO, SOC, SIS10K, and COME15K), our KepSalinst outperforms all previous models qualitatively and quantitatively.

Simultaneous Dimensionality Reduction and Classification via Dual Embedding Regularized Nonnegative Matrix Factorization.

Nonnegative matrix factorization (NMF) is a well-known paradigm for data representation. Traditional NMF-based classification methods first perform NMF or one of its variants on input data samples to obtain their low-dimensional representations, which are successively classified by means of a typical classifier [e.g., k -nearest neighbors (KNN) and support vector machine (SVM)]. Such a stepwise manner may overlook the dependency between the two processes, resulting in the compromise of the classification accuracy. In this paper, we elegantly unify the two processes by formulating a novel constrained optimization model, namely dual embedding regularized NMF (DENMF), which is semi-supervised. Our DENMF solution simultaneously finds the low-dimensional representations and assignment matrix via joint optimization for better classification. Specifically, input data samples are projected onto a couple of low-dimensional spaces (i.e., feature and label spaces), and locally linear embedding is employed to preserve the identical local geometric structure in different spaces. Moreover, we propose an alternating iteration algorithm to solve the resulting DENMF, whose convergence is theoretically proven. Experimental results over five benchmark datasets demonstrate that DENMF can achieve higher classification accuracy than state-of-the-art algorithms.

Using virtual reality to train emotional and social skills in children with autism spectrum disorder.

With emerging technology, computerised, internet-based and virtual reality (VR)-based treatment and training became increasingly popular. VR provides an immersive experience into a simulated environment. Autism spectrum disorder (ASD) is characterised by social communication deficits and repetitive behaviours. Children with ASD often require social skills training while VR provides a safe, controllable environment to practice skills repeatedly. The Centre for Innovative Applications of Internet and Multimedia Technologies (AIMTech Centre) at City University of Hong Kong developed a VR-enabled training program to examine its efficacy on emotional and social skills with six VR scenarios depicting the daily lives of typical children in Hong Kong. 94 children from mainstream primary schools in Hong Kong completed the study and 72 children were included in the analysis. Children from training group scored higher on emotion expression and regulation (M = 20.2, SD = 3.00) than before the training (M = 18.9, SD = 3.57, t(35) = -2.174, p = .037) and higher on social interaction and adaptation after the training (M = 21.8, SD = 2.99) than before training (M = 20.2, SD = 3.43, t(35) = -3.987, p < .0005). There was a statistically significant interaction between group and time on affective expressions, F(1, 70) = 5.223, p = .025, partial η2 = .069, and on social reciprocity, F(1, 70) = 7.769, p = .007, partial η2 = .100. Children were able to engage in VR training despite initial challenges with viewing goggles. Some children declined to participate due to scheduling reasons which may be minimised through the adoption of head-mounted displays as a portable, cost-effective device. VR seems to be a promising asset to traditional training and therapy while the importance trainers' or therapists' support has yet to be further investigated.