KNOWLEDGE-BASED REASONING NETWORK FOR OBJECT DETECTION

被引:5
|
作者
Zhang, Huigang [1 ]
Wang, Liuan [1 ]
Sun, Jun [1 ]
机构
[1] Fujitsu R&D Ctr, Beijing, Peoples R China
关键词
Object detection; commonsense knowledge; reasoning; GAT; COCO detection benchmark;
D O I
10.1109/ICIP42928.2021.9506228
中图分类号
TP18 [人工智能理论];
学科分类号
081104 ; 0812 ; 0835 ; 1405 ;
摘要
The mainstream object detection algorithms rely on recognizing object instances individually, but do not consider the high-level relationship among objects in context. This will inevitably lead to biased detection results, due to the lack of commonsense knowledge that humans often use to assist the task for object identification. In this paper, we present a novel reasoning module to endow the current detection systems with the power of commonsense knowledge. Specifically, we use graph attention network (GAT) to represent the knowledge among objects. The knowledge covers visual and semantic relations. Through the iterative update of GAT, the object features can be enriched. Experiments on the COCO detection benchmark indicate that our knowledge-based reasoning network has achieved consistent improvements upon various CNN detectors. We achieved 1.9 and 1.8 points higher Average Precision (AP) than Faster-RCNN and Mask-RCNN respectively, when using ResNet50-FPN as back-bone.
引用
收藏
页码:1579 / 1583
页数:5
相关论文
共 50 条
  • [11] Knowledge-based semantic reasoning for creativity
    Jing D.
    Tian Y.
    Zhang C.
    Yang C.
    Yang H.
    International Journal of Performability Engineering, 2020, 16 (05) : 800 - 810
  • [12] GENERIC TASKS IN KNOWLEDGE-BASED REASONING
    THIBAULT, RC
    IEEE EXPERT-INTELLIGENT SYSTEMS & THEIR APPLICATIONS, 1987, 2 (02): : 5 - 5
  • [13] Consequential reasoning in knowledge-based systems
    Hudson, DL
    Cohen, ME
    COMPUTERS AND THEIR APPLICATIONS, 2001, : 342 - 345
  • [14] A knowledge-based system for prototypical reasoning
    Lieto, Antonio
    Minieri, Andrea
    Piana, Alberto
    Radicioni, Daniele P.
    CONNECTION SCIENCE, 2015, 27 (02) : 137 - 152
  • [15] Impact Damage Detection in Aircraft Composites Using Knowledge-based Reasoning
    Mujica, L. E.
    Vehi, J.
    Staszewski, W.
    Worden, K.
    STRUCTURAL HEALTH MONITORING-AN INTERNATIONAL JOURNAL, 2008, 7 (03): : 215 - 230
  • [16] Object detection based on knowledge graph network
    Li, Jianping
    Tan, Guozhen
    Ke, Xiao
    Si, Huaiwei
    Peng, Yanfei
    APPLIED INTELLIGENCE, 2023, 53 (12) : 15045 - 15066
  • [17] Object detection based on knowledge graph network
    Jianping Li
    Guozhen Tan
    Xiao Ke
    Huaiwei Si
    Yanfei Peng
    Applied Intelligence, 2023, 53 : 15045 - 15066
  • [18] Knowledge-based methods for small object-detection in SAR images
    Greig, D
    Denny, M
    SAR IMAGE ANALYSIS, MODELING, AND TECHNIQUES V, 2003, 4883 : 121 - 130
  • [19] Few-Shot Object Detection Method Based on Knowledge Reasoning
    Wang, Jianwei
    Chen, Deyun
    ELECTRONICS, 2022, 11 (09)
  • [20] PROFESSIONAL VISION: FOCUS ON KNOWLEDGE-BASED REASONING
    Ulicna, Klara
    PROCEEDINGS OF THE 14TH INTERNATIONAL CONFERENCE EFFICIENCY AND RESPONSIBILITY IN EDUCATION 2017 (ERIE), 2017, : 482 - 489