English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

Sensor-based human activity segmentation and recognition are two importantand challenging problems in many real-world applications and they have drawnincreasing attention from the deep learning community in recent years. Most ofthe existing deep learning works were designed based on pre-segmented sensorstreams and they have treated activity segmentation and recognition as twoseparate tasks. In practice, performing data stream segmentation is verychallenging. We believe that both activity segmentation and recognition mayconvey unique information which can complement each other to improve theperformance of the two tasks. In this paper, we firstly proposes a newmultitask deep neural network to solve the two tasks simultaneously. Theproposed neural network adopts selective convolution and features multiscalewindows to segment activities of long or short time durations. First, multiplewindows of different scales are generated to center on each unit of the featuresequence. Then, the model is trained to predict, for each window, the activityclass and the offset to the true activity boundaries. Finally, overlappingwindows are filtered out by non-maximum suppression, and adjacent windows ofthe same activity are concatenated to complete the segmentation task. Extensiveexperiments were conducted on eight popular benchmarking datasets, and theresults show that our proposed method outperforms the state-of-the-art methodsboth for activity recognition and segmentation.

A Multi-Task Deep Learning Approach for Sensor-based Human Activity Recognition and Segmentation