arXiv Analytics

Sign in

arXiv:1810.03272 [cs.CV]AbstractReferencesReviewsResources

Light-Weight RefineNet for Real-Time Semantic Segmentation

Vladimir Nekrasov, Chunhua Shen, Ian Reid

Published 2018-10-08Version 1

We consider an important task of effective and efficient semantic image segmentation. In particular, we adapt a powerful semantic segmentation architecture, called RefineNet, into the more compact one, suitable even for tasks requiring real-time performance on high-resolution inputs. To this end, we identify computationally expensive blocks in the original setup, and propose two modifications aimed to decrease the number of parameters and floating point operations. By doing that, we achieve more than twofold model reduction, while keeping the performance levels almost intact. Our fastest model undergoes a significant speed-up boost from 20 FPS to 55 FPS on a generic GPU card on 512x512 inputs with solid 81.1% mean iou performance on the test set of PASCAL VOC, while our slowest model with 32 FPS (from original 17 FPS) shows 82.7% mean iou on the same dataset. Alternatively, we showcase that our approach is easily mixable with light-weight classification networks: we attain 79.2% mean iou on PASCAL VOC using a model that contains only 3.3M parameters and performs only 9.3B floating point operations.

Comments: Models are available here: https://github.com/drsleep/light-weight-refinenet, BMVC 2018
Categories: cs.CV
Related articles: Most relevant | Search more
arXiv:1806.02705 [cs.CV] (Published 2018-06-07)
Efficient semantic image segmentation with superpixel pooling
arXiv:1808.00897 [cs.CV] (Published 2018-08-02)
BiSeNet: Bilateral Segmentation Network for Real-time Semantic Segmentation
arXiv:1807.07466 [cs.CV] (Published 2018-07-19)
Guided Upsampling Network for Real-Time Semantic Segmentation