This paper presents the open-source eye tracking platform CrowdWatcher. It enables researchers to measure gaze location and user engagement in a crowdsourcing context through traditional RGB webcams. The proposed plat...
This paper presents the open-source eye tracking platform CrowdWatcher. It enables researchers to measure gaze location and user engagement in a crowdsourcing context through traditional RGB webcams. The proposed platform particularly advances the field of Quality of Experience (QoE) research, as it allows the experimenter to collect remotely and with very limited effort novel information from crowds of participants, such as their commitment towards a task, attention and decision-making processes. Two different experiments are described that were conducted to demonstrate the platform’s potential. The first experiment addresses the measurement of participants’ behavior while performing a movie selection task. Results show that the platform provides complementary information to traditional self-reported data by taking gaze analysis into account. This is of particular relevance, since in a crowdsourcing context decision processes and attention are difficult to assess, and there is often limited control over the engagement of the test user with the task. A second experiment is conducted in the scenario of a multimedia QoE test. Prediction accuracy is compared to a professional infrared eye tracker. While CrowdWatcher performs less well than the professional eye tracker, it is still able to collect valuable gaze information in the far more challenging environment of crowdsourcing. As an outlook to further application domains, the usage of the platform to measure user engagement allows participants who do not pay attention to the task to be identified.
Minimizing Gaussian curvature of meshes is fundamentally important for obtaining smooth and developable surfaces. However, there is a lack of computationally efficient and robust Gaussian curvature optimization method...
详细信息
In this paper, a kinematic motion planning algorithm for cooperative spatial payload manipulation is presented. A hierarchical approach is introduced to compute realtime collision-free motion plans for a formation of ...
详细信息
Dense pixelwise prediction such as semantic segmentation is an up-to-date challenge for deep convolutional neural networks (CNNs). Many state-of-the-art approaches either tackle the loss of high-resolution information...
详细信息
Most existing approaches to autonomous driving fall into one of two categories: Modular pipelines, that build an extensive model of the environment, and imitation learning approaches, that map images directly to contr...
详细信息
Background: Systems Medicine is a novel approach to medicine, that is, an interdisciplinary field that considers the human body as a system, composed of multiple parts and of complex relationships at multiple levels, ...
详细信息
The number of international benchmarking competitions is steadily increasing in various fields of machine learning (ML) research and practice. So far, however, little is known about the common practice as well as bott...
详细信息
We demonstrate improved performance in the classification of bioelectric data for use in systems such as robotic prosthesis control, by data fusion using low-cost electromyography (EMG) and electroencephalography (EEG...
We demonstrate improved performance in the classification of bioelectric data for use in systems such as robotic prosthesis control, by data fusion using low-cost electromyography (EMG) and electroencephalography (EEG) devices. Prosthetic limbs are typically controlled through EMG, and whilst there is a wealth of research into the use of EEG as part of a brain-computer interface (BCI) the cost of EEG equipment commonly prevents this approach from being adopted outside the lab. This study demonstrates as a proof-of-concept that multimodal classification can be achieved by using low-cost EMG and EEG devices in tandem, with statistical decision-level fusion, to a high degree of accuracy. We present multiple fusion methods, including those based on Jensen-Shannon divergence which had not previously been applied to this problem. We report accuracies of up to 99% when merging both signal modalities, improving on the best-case single-mode classification. We hence demonstrate the strengths of combining EMG and EEG in a multimodal classification system that could in future be leveraged as an alternative control mechanism for robotic prostheses.
暂无评论