Visualizing and Modifying Difficult Pixels in Cell Image Segmentation

Daisuke Matsuzuki, Kazuhiro Hotta

Abstract

In this paper, we visualize and modify difficult pixels to recognize for deep learning. In general, an image includes pixels that are easy or difficult to recognize. At the final layer, many deep learning methods use a softmax function to convert the outputs of network to probabilities. Pixels with small maximum probability are often difficult to recognize. We visualize those difficult pixels in a test image using the relationship between confidence and pixel-wise difficulty. By visualizing difficult pixels, we confirm the connection of cell membrane that could not be recognized by conventional method. We can connect the cell membrane by modifying difficult pixels. In experiments, we use cell image of mouse liver dataset including three classes; “cell membrane”, “cell nucleus” and “cytoplasm”. Our proposed method shows high recall score for “cell membrane”. We also confirmed the connection of cell membrane in qualitative evaluation.

Download


Paper Citation


in Harvard Style

Matsuzuki D. and Hotta K. (2020). Visualizing and Modifying Difficult Pixels in Cell Image Segmentation.In Proceedings of the 13th International Joint Conference on Biomedical Engineering Systems and Technologies - Volume 4: BIOSIGNALS, ISBN 978-989-758-398-8, pages 314-318. DOI: 10.5220/0009164203140318


in Bibtex Style

@conference{biosignals20,
author={Daisuke Matsuzuki and Kazuhiro Hotta},
title={Visualizing and Modifying Difficult Pixels in Cell Image Segmentation},
booktitle={Proceedings of the 13th International Joint Conference on Biomedical Engineering Systems and Technologies - Volume 4: BIOSIGNALS,},
year={2020},
pages={314-318},
publisher={SciTePress},
organization={INSTICC},
doi={10.5220/0009164203140318},
isbn={978-989-758-398-8},
}


in EndNote Style

TY - CONF

JO - Proceedings of the 13th International Joint Conference on Biomedical Engineering Systems and Technologies - Volume 4: BIOSIGNALS,
TI - Visualizing and Modifying Difficult Pixels in Cell Image Segmentation
SN - 978-989-758-398-8
AU - Matsuzuki D.
AU - Hotta K.
PY - 2020
SP - 314
EP - 318
DO - 10.5220/0009164203140318