상세정보
학술지
Three-stream Network with Context Convolution Module for Human-object Interaction Detection
- 발행일
- 202004
- 출처
- ETRI Journal, v.42 no.2, pp.230-238
- ISSN
- 1225-6463
- 출판사
- 한국전자통신연구원 (ETRI)
- 협약과제
- 19MH1500, 5G 기반의 스마트시티 서비스 개발 및 실증, 한미경
- 초록
- Human?뱋bject interaction (HOI) detection is a popular computer vision task that detects interactions between humans and objects. This task can be useful in many applications that require a deeper understanding of semantic scenes. Current HOI detection networks typically consist of a feature extractor followed by detection layers comprising small filters (eg, 1혻×혻1 or 3혻×혻3). Although small filters can capture local spatial features with a few parameters, they fail to capture larger context information relevant for recognizing interactions between humans and distant objects owing to their small receptive regions. Hence, we herein propose a three-stream HOI detection network that employs a context convolution module (CCM) in each stream branch. The CCM can capture larger contexts from input feature maps by adopting combinations of large separable convolution layers and residual-based convolution layers without increasing the number of parameters by using fewer large separable filters. We evaluate our HOI detection method using two benchmark datasets, V-COCO and HICO-DET, and demonstrate its state-of-the-art performance.
- KSP 제안 키워드
- Art performance, Benchmark datasets, Computer Vision(CV), Context Information, Detection Method, Feature Map, Human-Object Interaction, Interaction detection, Separable Filters, Stream network, feature extractor