Abstract
The study of motion offers a unique perspective on human physiology and individuality. Motion represents progressions of simultaneous maneuvers of body parts. Analyzing motions provides a glimpse into the nuanced characteristics of movements. Recent advances in deep-learning techniques have shown remarkable success in action-related research. However, these researches mainly focus on atomic actions with limited gestures, where mapping spatiotemporal dependencies is less challenging. Current methodologies face ongoing difficulties for actions with higher complexities. It is crucial to extract discriminative motion dynamics for detailed comprehension of actions. Therefore, this dissertation aims to study motions from various perspectives to address thegaps.
Original language | English |
---|---|
Qualification | Doctor of Philosophy |
Awarding Institution |
|
Supervisors/Advisors |
|
Award date | 30 Aug 2024 |
DOIs | |
Publication status | Unpublished - 2024 |