NOTE-RCNN: NOise Tolerant Ensemble RCNN for Semi-Supervised Object Detection.
2019 IEEE/CVF International Conference on Computer Vision (ICCV)（2019）
关键词
NOTE-RCNN,labeling cost,training modern object detectors,expensive bounding box annotations,semisupervised object detection formulation,seed box level annotations,image level annotations,training-mining framework,weakly supervised object detection tasks,mining process,labelling noises,false negatives,false positives,inaccurate boundaries,standard object detectors,novel noise tolerant ensemble RCNN,noisy labels,standard Faster RCNN,classification heads,distillation head,mining precision,negative sample loss,box predictor,false negative labels,training box regression head,seed annotations,mined bounding boxes,detection accuracy,training steps
AI 理解论文
溯源树
样例
生成溯源树，研究论文发展脉络
Chat Paper
正在生成论文摘要