—Data size is the bottleneck for developing deep saliency models, because collecting eye-movement data is very time-consuming and expensive. Most of current studies on human attention and saliency modeling have used ...
详细信息
—Data size is the bottleneck for developing deep saliency models, because collecting eye-movement data is very time-consuming and expensive. Most of current studies on human attention and saliency modeling have used high-quality stereotype stimuli. In real world, however, captured images undergo various types of transformations. Can we use these transformations to augment existing saliency datasets? Here, we first create a novel saliency dataset including fixations of 10 observers over 1900 images degraded by 19 types of transformations. Second, by analyzing eye movements, we find that observers look at different locations over transformed versus original images. Third, we utilize the new data over transformed images, called data augmentation transformation (DAT), to train deep saliency models. We find that label-preserving DATs with negligible impact on human gaze boost saliency prediction, whereas some other DATs that severely impact human gaze degrade the performance. These label-preserving valid augmentation transformations provide a solution to enlarge existing saliency datasets. Finally, we introduce a novel saliency model based on generative adversarial network (dubbed GazeGAN). A modified U-Net is proposed as the generator of the GazeGAN, which combines classic "skip connections" with a novel "center-surround connection (CSC)", in order to leverage multi-level features. We also propose a histogram loss based on Alternative Chi-Square Distance (ACS HistLoss) to refine the saliency map in terms of luminance distribution. Extensive experiments and comparisons over 3 datasets indicate that GazeGAN achieves the best performance in terms of popular saliency evaluation metrics, and is more robust to various perturbations. We also provide a comprehensive quantitative comparison of 22 state-of-the-art saliency models on distorted images, which contributes a robustness benchmark for saliency community. Our code and data are available at: https://***/CZHQuality/
Space division multiplexing, in conjunction with wavelength division multiplexing, is a powerful mechanism that increases spatial channels significantly in a single optical fiber. Spiral Phased-Laguerre-Gaussian (SP-L...
Space division multiplexing, in conjunction with wavelength division multiplexing, is a powerful mechanism that increases spatial channels significantly in a single optical fiber. Spiral Phased-Laguerre-Gaussian (SP-LG) modes indicate that they effectively reduce mode coupling, thereby increasing the channel response and performance positively. However, mode-coupling influence optical fiber system leads to inter-symbol interference (ISI) between the channels and reduces both capacity and distance. In this paper, we apply the SP-LG modes in a space wavelength division multiplexing over a few-mode fiber system. The results demonstrate significant improvements in reducing channels effects and ISI. The performance evaluations are based on the bit-error-rate (BER), eye diagram and spectrum analyzer.
Home-based exercise programs can improve physical functioning and health status of elderly people. Successful implementation of exercise interventions for older people presents major challenges and supporting elderly ...
详细信息
Home-based exercise programs can improve physical functioning and health status of elderly people. Successful implementation of exercise interventions for older people presents major challenges and supporting elderly people properly while doing their home-based exercises is essential for training success. We developed a tablet-based system-ActiveLifestyle-that offers older adults a home-based strength-balance training program with incorporated motivation strategies and support features. Objective: The goal was to compare 3 different home-based training programs with respect to their effect on measures of gait quality and physical performance through planned comparisons between (1) tablet-based and brochure-based interventions, (2) individual and social motivation strategies, and (3) active and inactive participants. Methods: A total of 44 autonomous-living elderly people (mean 75, SD 6 years) were assigned to 3 training groups: social (tablet guided, n=14), individual (tablet guided, n=13), and brochure (brochure guided, n=17). All groups joined a 12-week progressive home-based strength-balance training program. Outcome measures were gait performance under single and dual task conditions, dual task costs of walking, falls efficacy, and physical performance as measured by the Short Physical Performance Battery (SPPB). Furthermore, active (≥75% program compliance) and inactive (<75% program compliance) individuals were compared based on their characteristics and outcome measures. Results: The tablet groups showed significant improvements in single and dual task walking, whereas there were no significant changes observable in the brochure group. Between-groups comparisons revealed significant differences for gait velocity (U=138.5;P=.03, r=.33) and cadence (U=138.5, P=.03 r=.34) during dual task walking at preferred speed in favor of the tablet groups. The brochure group had more inactive participants, but this did not reach statistical significance (U=167, P=.06, r=.2
This paper deals with optical code division multiple access (optical-CDMA) and provides a general analysis of multi-mode fiber (MMF). The purpose of this work is to classify the works in the literature that are relate...
This paper deals with optical code division multiple access (optical-CDMA) and provides a general analysis of multi-mode fiber (MMF). The purpose of this work is to classify the works in the literature that are related to Optical-CDMA based on MMF and to find the limitation of the researches done in this field. Several challenges that occur in the medium such as multiple-access interference (MAI), pulse dispersion (PD) and nonlinearity in optical fibers deteriorate system performances and become a major performance-limiting factor.
Choosing an appropriate optimization algorithm is essential to achieving success in optimization challenges. Here we present a new evolutionary algorithm structure that utilizes a reinforcement learning-based agent ai...
详细信息
Multimodal manipulations (also known as audiovisual deepfakes) make it difficult for unimodal deepfake detectors to detect forgeries in multimedia content. To avoid the spread of false propaganda and fake news, timely...
详细信息
In this paper we propose the use of an adaptive multivariate approach to define time-varying multivariate estimators based on the Directed Transfer Function (DTF) and the Partial Directed Coherence (PDC). DTF and PDC ...
详细信息
In this paper we propose the use of an adaptive multivariate approach to define time-varying multivariate estimators based on the Directed Transfer Function (DTF) and the Partial Directed Coherence (PDC). DTF and PDC are frequency-domain estimators that are able to describe interactions between cortical areas in terms of the concept of Granger causality. Time-varying DTF and PDC were obtained by the adaptive recursive fit of an MVAR model with time-dependent parameters, by means of a generalized recursive least-square (RLS) algorithm, taking into consideration a set of EEG epochs. Such estimators are able to follow rapid changes in the connectivity between cortical areas during an experimental task. We provide an application to the cortical estimations obtained from high resolution EEG data, recorded from a group of healthy subject during a combined foot-lips movement, and present the time-varying connectivity patterns resulting from the application of both DTF and PDC. Two different cortical networks were detected, one constant across the task and the other evolving during the preparation of the joint movement.
Melanoma is the most deadly form of skin cancer. Tracking the evolution of nevi and detecting new lesions across the body is essential for the early detection of melanoma. Despite prior work on longitudinal tracking o...
详细信息
暂无评论