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

One of the practical choices for making a lightweight semantic segmentationmodel is to combine a depth-wise separable convolution with a dilatedconvolution. However, the simple combination of these two methods results in anover-simplified operation which causes severe performance degradation due toloss of information contained in the feature map. To resolve this problem, wepropose a new block called Concentrated-Comprehensive Convolution (C3) whichapplies the asymmetric convolutions before the depth-wise separable dilatedconvolution to compensate for the information loss due to dilated convolution.The C3 block consists of a concentration stage and a comprehensive convolutionstage. The first stage uses two depth-wise asymmetric convolutions forcompressed information from the neighboring pixels to alleviate the informationloss. The second stage increases the receptive field by using a depth-wiseseparable dilated convolution from the feature map of the first stage. Weapplied the C3 block to various segmentation frameworks (ESPNet, DRN, ERFNet,ENet) for proving the beneficial properties of our proposed method.Experimental results show that the proposed method preserves the originalaccuracies on Cityscapes dataset while reducing the complexity. Furthermore, wemodified ESPNet to achieve about 2% better performance while reducing thenumber of parameters by half and the number of FLOPs by 35% compared with theoriginal ESPNet. Finally, experiments on ImageNet classification task show thatC3 block can successfully replace dilated convolutions.

C3: Concentrated-Comprehensive Convolution and its application to semantic segmentation