scholarly journals Multimodal Attention Network for Trauma Activity Recognition from Spoken Language and Environmental Sound

Author(s):  
Yue Gu ◽  
Ruiyu Zhang ◽  
Xinwei Zhao ◽  
Shuhong Chen ◽  
Jalal Abdulbaqi ◽  
...  
2019 ◽  
Vol 1 (2) ◽  
pp. 100-109
Author(s):  
Hepnyi Samosir

Even though listening is often categorized as a passive activity, it is not a simple process as it may seem. There is an active process because in it. Listeners must focus on the spoken language and actively draw what message is being delivered. There are some factors that make a listener may need a lot of work to comprehend a spoken language such as colloquialism, accent, intonation, hesitation, word or phrase clusters, and content. Preliminary data gained from pre-observation at STMIK Prabumulih showed that the third semester students found themselves hard to acquire information from spoken language, hence it is necessary to find out the causes so that make an effort to develop their listening skills. The method used is descriptive research design. The result reveals that students encounter various kinds of listening problems in learning comprehension such as unfamiliar words, the length and of the spoken text, speed rate, lack of concentration and pronunciation, and the noises around come from both the background sound of the spoken language recording and environmental sound such as sound from outside classroom. Keywords: listening; comprehension; listening comprehension problems; descriptive


2019 ◽  
Vol 11 (21) ◽  
pp. 2584 ◽  
Author(s):  
Yuan He ◽  
Xinyu Li ◽  
Xiaojun Jing

Short-range radar has become one of the latest sensor technologies for the Internet of Things (IoT), and it plays an increasingly vital role in IoT applications. As the essential task for various smart-sensing applications, radar-based human activity recognition and person identification have received more attention due to radar’s robustness to the environment and low power consumption. Activity recognition and person identification are generally treated as separate problems. However, designing different networks for these two tasks brings a high computational complexity and wastes of resources to some extent. Furthermore, there are some correlations in activity recognition and person identification tasks. In this work, we propose a multiscale residual attention network (MRA-Net) for joint activity recognition and person identification with radar micro-Doppler signatures. A fine-grained loss weight learning (FLWL) mechanism is presented for elaborating a multitask loss to optimize MRA-Net. In addition, we construct a new radar micro-Doppler dataset with dual labels of activity and identity. With the proposed model trained on this dataset, we demonstrate that our method achieves the state-of-the-art performance in both radar-based activity recognition and person identification tasks. The impact of the FLWL mechanism was further investigated, and ablation studies of the efficacy of each component in MRA-Net were also conducted.


Sign in / Sign up

Export Citation Format

Share Document