English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

Feature pyramid architecture has been broadly adopted in object detection andsegmentation to deal with multi-scale problem. However, in this paper we showthat the capacity of the architecture has not been fully explored due to theinadequate utilization of the supervision information. Such insufficientutilization is caused by the supervision signal degradation in backpropagation. Thus inspired, we propose a dually supervised method, named duallysupervised FPN (DSFPN), to enhance the supervision signal when training thefeature pyramid network (FPN). In particular, DSFPN is constructed by attachingextra prediction (i.e., detection or segmentation) heads to the bottom-upsubnet of FPN. Hence, the features can be optimized by the additional headsbefore being forwarded to subsequent networks. Further, the auxiliary heads canserve as a regularization term to facilitate the model training. In addition,to strengthen the capability of the detection heads in DSFPN for handling twoinhomogeneous tasks, i.e., classification and regression, the originally sharedhidden feature space is separated by decoupling classification and regressionsubnets. To demonstrate the generalizability, effectiveness, and efficiency ofthe proposed method, DSFPN is integrated into four representative detectors(Faster RCNN, Mask RCNN, Cascade RCNN, and Cascade Mask RCNN) and assessed onthe MS COCO dataset. Promising precision improvement, state-of-the-artperformance, and negligible additional computational cost are demonstratedthrough extensive experiments. Code will be provided.

Dually Supervised Feature Pyramid for Object Detection and Segmentation