Abstract:
Understanding human-object interaction (HOI) requires detection. Interactiveness Knowledge—whether humans and objects interact—is examined in this paper. Interactiveness knowledge can be learned across HOI datasets and bridge diverse HOI category settings. We use an Interactiveness Network to learn general interactiveness from multiple HOI datasets and perform Non-Interaction Suppression before HOI classification in inference. Due to its generalization, interactiveness network is a transferable knowledge learner that can work with any HOI detection model to achieve desired results. We learn hierarchical paradigm interactiveness using human instance and body part features. A consistency task guides learning and extracts deeper interactive visual clues. We test the method on HICO-DET, V-COCO, and HAKE-HOI. Our learned interactiveness outperforms state-of-the-art HOI detection methods, proving its efficacy and flexibility.
Note: Please discuss with our team before submitting this abstract to the college. This Abstract or Synopsis varies based on student project requirements.
Did you like this final year project?
To download this project Code with thesis report and project training... Click Here
