no code implementations • 23 Aug 2023 • Sota Kato, Kazuhiro Hotta, Yuhki Hatakeyama, Yoshinori Konishi
Lite-HRNet Plus achieves two improvements: a novel fusion block based on a channel attention and a novel output module with less computational intensity using multi-resolution feature maps.
1 code implementation • 15 Jun 2023 • Sota Kato, Kazuhiro Hotta
Although, by using LDAM loss, it is possible to obtain large margins for the minority classes and small margins for the majority classes, the relevance to a large margin, which is included in the original softmax cross entropy loss, is not be clarified yet.
no code implementations • 21 Apr 2023 • Taiji Kurami, Takuya Ishikawa, Kazuhiro Hotta
However, the images of pancreatic tissue fragments used in this study cannot be successfully classified by processing the entire image because the pancreatic tissue fragments are only a part of the image.
1 code implementation • 17 Apr 2023 • Sota Kato, Kazuhiro Hotta
Semantic segmentation of microscopic cell images using deep learning is an important technique, however, it requires a large number of images and ground truth labels for training.
no code implementations • 14 Oct 2022 • Shinji Yamada, Satoshi Kamiya, Kazuhiro Hotta
To improve the accuracy of STPM, this work uses a student network, as in generative models, to reconstruct normal features.
Ranked #32 on Anomaly Detection on MVTec AD
1 code implementation • 16 Jul 2022 • Sota Kato, Kazuhiro Hotta
Based on the t-vMF similarity, our proposed Dice loss is formulated in a more compact similarity loss function than the original Dice loss.
Ranked #1 on Medical Image Segmentation on CVC-ClinicDB
no code implementations • 20 Mar 2022 • Hiroki Tsuda, Kazuhiro Hotta
Finally, we propose an Adversarial Mutual Leakage Network (AML-Net) that mutually leaks the information each other between the generator and the discriminator.
no code implementations • 3 Dec 2021 • Ryouichi Furukawa, Kazuhiro Hotta
The proposed method recovers positional information by emphasizing the similarity between decoder's feature maps with superior semantic information and encoder's feature maps with superior positional information.
no code implementations • 30 Nov 2021 • Shinji Yamada, Kazuhiro Hotta
The other student-teacher network has a role to reconstruct the features of normal products.
no code implementations • 30 Aug 2021 • Sota Kato, Kazuhiro Hotta
We propose an automatic preprocessing and ensemble learning for segmentation of cell images with low quality.
no code implementations • 6 Jul 2021 • Sota Kato, Kazuhiro Hotta
Unlike CE loss, MSE loss is possible to equalize the number of back propagation for all classes and to learn the feature space considering the relationships between classes as metric learning.
no code implementations • 20 Jan 2021 • Ryota Ikedo, Kazuhiro Hotta
By sharing feature maps, one of two networks can obtain the information that cannot be obtained by a single network.
no code implementations • 20 Jan 2021 • Takamasa Ando, Kazuhiro Hotta
Although loss function is optimized in training deep neural network, far layers from the layers for computing loss function are difficult to train.
no code implementations • 14 Aug 2020 • Hiroki Tsuda, Eisuke Shibuya, Kazuhiro Hotta
In this paper, we address cell image segmentation task by Feedback Attention mechanism like feedback processing.
no code implementations • 30 Apr 2020 • Eisuke Shibuya, Kazuhiro Hotta
By using Convolutional LSTM, the features in the second round are extracted based on the features acquired in the first round.
no code implementations • 28 Mar 2017 • Shohei Kumagai, Kazuhiro Hotta, Takio Kurita
In this paper, we propose to predict the number of targets using multiple CNNs specialized to a specific appearance, and those CNNs are adaptively selected according to the appearance of a test image.