Superpixel-based semantic segmentation trained by statistical process control

Hyojin Park, Jisoo Jeong, Youngjoon Yoo and Nojun Kwak

Abstract

Semantic segmentation, like other fields of computer vision, has seen a remarkable performance advance by the use of deep convolution neural networks. However, considering that neighboring pixels are heavily dependent on each other, both learning and testing of these methods have a lot of redundant operations. To resolve this problem, the proposed network is trained and tested with only 0.37% of total pixels by superpixel-based sampling and largely reduced the complexity of upsampling calculation. The hypercolumn feature maps are constructed by pyramid module in combination with the convolution layers of the base network. Since the proposed method uses a very small number of sampled pixels, the end-to-end learning of the entire network is difficult with a common learning rate for all the layers. In order to resolve this problem, the learning rate after sampling is controlled by statistical process control (SPC) of gradients in each layer.

Session

Posters

Files

Paper (PDF)

DOI

10.5244/C.31.78
https://dx.doi.org/10.5244/C.31.78

Citation

Bibtex

            @inproceedings{BMVC2017_78,
                title={Superpixel-based semantic segmentation trained by statistical process control},
                author={Hyojin Park, Jisoo Jeong, Youngjoon Yoo and Nojun Kwak},
                year={2017},
                month={September},
                pages={78.1-78.13},
                articleno={78},
                numpages={13},
                booktitle={Proceedings of the British Machine Vision Conference (BMVC)},
                publisher={BMVA Press},
                editor={Tae-Kyun Kim, Stefanos Zafeiriou, Gabriel Brostow and Krystian Mikolajczyk},
                doi={10.5244/C.31.78},
                isbn={1-901725-60-X},
                url={https://dx.doi.org/10.5244/C.31.78}
            }