Faculty Profiles - ODA, Masahiro

Continual pretraining for enhanced multi-organ segmentation from CT images

Yang, YQ; Shen, C; Tang, YC; Roth, HR; Oda, M; Hayashi, Y; Misawa, K; Mori, K

MEDICAL IMAGING 2024: IMAGE PROCESSING Vol. 12926 2024

Publisher：Progress in Biomedical Optics and Imaging - Proceedings of SPIE

Self-supervised pretraining has shown great performance in improving the accuracy of downstream tasks. Although pretraining on a large dataset improves performances, it becomes challenging to further optimize the model by solely enlarging the dataset. In contrast, additional adaptation of pretrained models to the target domain has shown promise in NLP. Inspired by the success of continual pretraining, we investigated the efficacy of adapting the target domain dataset to a pretrained model in medical imaging, particularly in the context of segmentation. We present a study based on a self-supervised pretraining framework using the SwinUNETR backbone. In this study, we improved the generalizability of the self-supervised pretraining by adapting a foundational model pretrained on 5k CT volumes to data of the downstream segmentation task. In detail, we employed 385 abdominal CT volumes for the continual task-adaptive pretraining and 24 abdominal CT volumes for the downstream segmentation task, all sourced from the same dataset. Additionally, we conducted comparative experiments to demonstrate the benefits of this task-adapting pretraining approach. Our method has shown that continual pretraining helps to improve the performances, achieving an average Dice score for 10-class organ segmentation of 87.8%.

DOI： 10.1117/12.3006630

Intestine Segmentation from CT Volume based on Bidirectional Teaching

An, Q; Oda, H; Hayashi, Y; Kitasaka, T; Hinoki, A; Uchida, H; Suzuki, K; Takimoto, A; Oda, M; Mori, K

MEDICAL IMAGING 2024: IMAGE PROCESSING Vol. 12926 2024

Publisher：Progress in Biomedical Optics and Imaging - Proceedings of SPIE

This paper proposes an intestine segmentation method to segment intestines from CT volumes for helping clinicians diagnose intestine obstruction. For large-scale labeled datasets, fully-supervised methods have shown superior results. However, medical image segmentation is usually difficult to achieve accurate prediction due to the limited number of labeled data available for training. To address this challenge, we introduce a novel multi-view symmetrical network (MVS-Net) for intestine segmentation and incorporate bidirectional teaching to utilize unlabeled datasets. Specifically, we design the MVS-Net, which can use different sizes of convolution kernels instead of a fixed kernel size, enabling the network to capture multi-scale features from images’ different perceptual fields and ensure segmentation accuracy. Additionally, the pseudo-labels are generated by bidirectional teaching, which can make the network captures semantic information from large-scale unlabeled data for increasing the training data. We repeated the experiment five times, and used the averaged result on the intestines dataset to represent the segmentation accuracy of the proposed method. The experimental results showed the average Dice was 78.86%, the average recall 84.50%, and the average precision 75.94%, respectively.

DOI： 10.1117/12.3006623

HTSeg: Hybrid Two-Stage Segmentation Framework for Intestine Segmentation from CT Volumes

An Q., Oda H., Hayashi Y., Kitasaka T., Takimoto A., Hinoki A., Uchida H., Suzuki K., Oda M., Mori K.

Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) Vol. 15196 LNCS page： 32 - 41 2024

Publisher：Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

This paper proposes a semi-supervised intestine segmentation method from CT volumes. Our method can use densely and sparsely annotated CT volumes for training to reduce the labor of manually annotating intestines. The proposed Hybrid Two-stage Segmentation (HTSeg) framework consists of two networks, a 2D swin-transformer-based network as the first stage and a 3D network as the second stage. In the first stage, we use 6964 labeled CT slices to train the 2D Swin U-Net. The trained 2D Swin U-Net is used to generate pseudo-labels for sparse annotation data. In the second stage, we use sparsely annotated datasets with pseudo-labels and densely annotated datasets to train a 3D multi-view symmetrical network (MVSNet). Experimental results showed that the Dice score of the proposed method was 74.70%, which was 1.03% higher than just using MVSNet. Compared with the other four previous methods (3D U-Net, CPS, EM, MT), the proposed method produced competitive segmentation performance. The code can be found at: https://github.com/MoriLabNU/semi-pseudo-labels.

DOI： 10.1007/978-3-031-73083-2_4

Performance Improvement for Medical Image Classification Model by using Gradient-based Analytical Feature Selection

Toda, R; Itoh, H; Oda, M; Hayashi, Y; Otake, Y; Hashimoto, M; Akashi, T; Aoki, S; Mori, K

COMPUTER-AIDED DIAGNOSIS, MEDICAL IMAGING 2024 Vol. 12927 2024

Publisher：Progress in Biomedical Optics and Imaging - Proceedings of SPIE

This paper presents a gradient-based analytical method for improving medical image classification. The automated classification of diseases is important in computer-aided diagnosis. In addition to accurate classification, its explainability is an essential factor toward its practical application. A gradient-based visual explanation provides the explainability of a model of an convolutional neural network (CNN) by indicating important patterns in an input image. Most studies use this explanation to assess CNN’s validity in a qualitative manner. On the other hand, in addition to a model’s explainability, our motivation is to utilize the visual-explanation methods to enhance the classification accuracy of a CNN model. In this study, we propose a weight-analysis-based method to improve the classification accuracy of a trained-CNN model without additional training. The proposed method selects important patterns based on a gradient-based weight analysis of a middle layer in a trained model and suppresses irrelevant patterns in the extracted features for the classification. We applied our analytical method to a convolutional and a global-average-pooling layers in a CNN, which classifies a chest CT volume into COVID-19 typical and non-typical cases. As shown in classification results on 302 testing cases, our method improved the accuracy of the COVID-19 classification.

DOI： 10.1117/12.3006620

A Bayesian Approach to Weakly-Supervised Laparoscopic Image Segmentation

Zheng, Z; Hayashi, Y; Oda, M; Kitasaka, T; Mori, K

MEDICAL IMAGE COMPUTING AND COMPUTER ASSISTED INTERVENTION - MICCAI 2024, PT VI Vol. 15006 page： 14 - 24 2024

Smart: Stoma Measurement, Analysis, Report Tool for Microscope Image and Its Application in Plant Phenotyping

Wang C., Sato K., Hayashi Y., Oda M., Ishimaru Y., Uozumi N., Mori K.

International Conference on Wavelet Analysis and Pattern Recognition 2024

Publisher：International Conference on Wavelet Analysis and Pattern Recognition

This manuscript describes a deep learning-based algorithm for inferring stomatal phenotypes from microscope images. Botanists spray compounds on leaf surfaces and observe their state under a microscope to study the effects of different compounds on stomatal opening and closing. We propose a stomatal orientation-based method that uses stomatal orientation to guide stomatal measurements. This method has three modules: a stomata detection module that locates the stoma region and orientation using deep learning-based object detection. A stoma segmentation module that segments the aperture, guard cell, and thick inner wall from the stoma ROI image. And a phenotype quantification module that calculates the phenotype parameters by analyzing the mask image. The experimental results show that the proposed method can resolve stomata with high accuracy (the average R2 of the previous method is 0.66, and the proposed method is 0.96). For the development of the community, we will release the algorithm and tool involved in this article in GitHub.

DOI： 10.1109/ICWAPR63074.2024.10870510

Semi-supervised Tubular Structure Segmentation with Cross Geometry and Hausdorff Distance Consistency

Zhu, RY; Oda, M; Hayashi, Y; Kitasaka, T; Mori, K

MEDICAL IMAGE COMPUTING AND COMPUTER ASSISTED INTERVENTION - MICCAI 2024, PT VIII Vol. 15008 page： 612 - 622 2024

Publisher：Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

This study introduces a novel semi-supervised method for 3D segmentation of tubular structures. Complete and automated segmentation of complex tubular structures in medical imaging remains a challenging task. Traditional supervised deep learning methods often demand a tremendous number of annotated data to train the deep model, with the high cost and difficulty of obtaining annotations. To address this, a semi-supervised approach could be a viable solution. Segmenting complex tubular structures with limited annotated data remains a formidable challenge. Many semi-supervised techniques rely on pseudo-labeling, which involves generating labels for unlabeled images based on predictions from a model trained on labeled data. Besides, several semi-supervised learning methods are proposed based on data-level consistency, which enforces consistent predictions by applying perturbations to input images. However, these methods tend to overlook the geometric shape characteristics of the segmentation targets. In our research, we introduce a task-level consistency learning approach that incorporates cross geometry consistency and the Hausdorff distance consistency, taking advantage of the geometric shape properties of both labeled and unlabeled data. Our deep learning model generates both a segmentation map and a distance transform map. By applying the proposed consistency, we ensure that the geometric shapes in both maps align closely, thereby enhancing the accuracy and performance of tubular structure segmentation. We tested our method on airway segmentation in 3D CT scans, where it outperformed the recent state-of-the-art methods, showing an 88.4% tree length detected rate, 82.8% branch detected rate, and 89.7% precision rate.

DOI： 10.1007/978-3-031-72111-3_58

Leaf Navigated Plant Growth Measurement for Automated Plant Cultivation System

Wang C., Ukai Y., Hayashi Y., Oda M., Sat Y., Mori K.

International Conference on Wavelet Analysis and Pattern Recognition 2024

Publisher：International Conference on Wavelet Analysis and Pattern Recognition

This manuscript describes a workflow for automatically measuring plant growth from scanned images. To explore the different conditions of plant growth, botanists sow seeds in trays and use imaging devices such as scanners or cameras to record the daily growth of plants. Previous methods behave poorly in root measuring. To design an automated plant growth measurement system, we propose a leaf-root distance-based plant organ measurement method that finds the paths of the root by minimizing the distance energy from the root endpoints to the leaves. This method contains three modules. The first is a deep learning-based plant detection module to detect individual plants' regions of interest (ROI). An organ mask generation module that uses Otsu-thresholding to segment root and leaf. A plant organ analysis module to generate plant phenotypes (e.g., length of primary and lateral roots) from mask images. The experimental results showed that the proposed method had a higher measurement accuracy than previous methods.

DOI： 10.1109/ICWAPR63074.2024.10870512

SGSR: style-subnets-assisted generative latent bank for large-factor super-resolution with registered medical image dataset. International journal

Tong Zheng, Hirohisa Oda, Yuichiro Hayashi, Shota Nakamura, Masaki Mori, Hirotsugu Takabatake, Hiroshi Natori, Masahiro Oda, Kensaku Mori

International journal of computer assisted radiology and surgery 2023.12

Language：English Publishing type：Research paper (scientific journal)

PURPOSE: We propose a large-factor super-resolution (SR) method for performing SR on registered medical image datasets. Conventional SR approaches use low-resolution (LR) and high-resolution (HR) image pairs to train a deep convolutional neural network (DCN). However, LR-HR images in medical imaging are commonly acquired from different imaging devices, and acquiring LR-HR image pairs needs registration. Registered LR-HR images have registration errors inevitably. Using LR-HR images with registration error for training an SR DCN causes collapsed SR results. To address these challenges, we introduce a novel SR approach designed specifically for registered LR-HR medical images. METHODS: We propose style-subnets-assisted generative latent bank for large-factor super-resolution (SGSR) trained with registered medical image datasets. Pre-trained generative models named generative latent bank (GLB), which stores rich image priors, can be applied in SR to generate realistic and faithful images. We improve GLB by newly introducing style-subnets-assisted GLB (S-GLB). We also propose a novel inter-uncertainty loss to boost our method's performance. Introducing more spatial information by inputting adjacent slices further improved the results. RESULTS: SGSR outperforms state-of-the-art (SOTA) supervised SR methods qualitatively and quantitatively on multiple datasets. SGSR achieved higher reconstruction accuracy than recently supervised baselines by increasing peak signal-to-noise ratio from 32.628 to 34.206 dB. CONCLUSION: SGSR performs large-factor SR while given a registered LR-HR medical image dataset with registration error for training. SGSR's results have both realistic textures and accurate anatomical structures due to favorable quantitative and qualitative results. Experiments on multiple datasets demonstrated SGSR's superiority over other SOTA methods. SR medical images generated by SGSR are expected to improve the accuracy of pre-surgery diagnosis and reduce patient burden.

DOI： 10.1007/s11548-023-03037-3

Automated Detection of the Thoracic Ossification of the Posterior Longitudinal Ligament Using Deep Learning and Plain Radiographs

researchmap

Ito, S; Nakashima, H; Segi, N; Ouchida, J; Oda, M; Yamauchi, I; Oishi, R; Miyairi, Y; Mori, K; Imagama, S

BIOMED RESEARCH INTERNATIONAL Vol. 2023 page： 8495937 2023.11

Language：English Publisher：BioMed Research International

Ossification of the ligaments progresses slowly in the initial stages, and most patients are unaware of the disease until obvious myelopathy symptoms appear. Consequently, treatment and clinical outcomes are not satisfactory. This study is aimed at developing an automated system for the detection of the thoracic ossification of the posterior longitudinal ligament (OPLL) using deep learning and plain radiography. We retrospectively reviewed the data of 146 patients with thoracic OPLL and 150 control cases without thoracic OPLL. Plain lateral thoracic radiographs were used for object detection, training, and validation. Thereafter, an object detection system was developed, and its accuracy was calculated. The performance of the proposed system was compared with that of two spine surgeons. The accuracy of the proposed object detection model based on plain lateral thoracic radiographs was 83.4%, whereas the accuracies of spine surgeons 1 and 2 were 80.4% and 77.4%, respectively. Our findings indicate that our automated system, which uses a deep learning-based method based on plain radiographs, can accurately detect thoracic OPLL. This system has the potential to improve the diagnostic accuracy of thoracic OPLL.

DOI： 10.1155/2023/8495937

Automated Detection and Diagnosis of Spinal Schwannomas and Meningiomas Using Deep Learning and Magnetic Resonance Imaging

ラベル拡張法によるマイクロCT像中の小葉間隔壁の抽出

深井大輔, 小田紘久, 林雄一郎, 鄭通, 中村彰太, 小田昌宏, 森健策

日本コンピュータ外科学会誌 Vol. 25 ( 3 ) page： 259 - 259 2023.11

Ito Sadayuki, Nakashima Hiroaki, Segi Naoki, Ouchida Jun, Oda Masahiro, Yamauchi Ippei, Oishi Ryotaro, Miyairi Yuichi, Mori Kensaku, Imagama Shiro

JOURNAL OF CLINICAL MEDICINE Vol. 12 ( 15 ) 2023.8

Language：English Publisher：Journal of Clinical Medicine

Spinal cord tumors are infrequently identified spinal diseases that are often difficult to diagnose even with magnetic resonance imaging (MRI) findings. To minimize the probability of overlooking these tumors and improve diagnostic accuracy, an automatic diagnostic system is needed. We aimed to develop an automated system for detecting and diagnosing spinal schwannomas and meningiomas based on deep learning using You Only Look Once (YOLO) version 4 and MRI. In this retrospective diagnostic accuracy study, the data of 50 patients with spinal schwannomas, 45 patients with meningiomas, and 100 control cases were reviewed, respectively. Sagittal T1-weighted (T1W) and T2-weighted (T2W) images were used for object detection, classification, training, and validation. The object detection and diagnosis system was developed using YOLO version 4. The accuracies of the proposed object detections based on T1W, T2W, and T1W + T2W images were 84.8%, 90.3%, and 93.8%, respectively. The accuracies of the object detection for two spine surgeons were 88.9% and 90.1%, respectively. The accuracies of the proposed diagnoses based on T1W, T2W, and T1W + T2W images were 76.4%, 83.3%, and 84.1%, respectively. The accuracies of the diagnosis for two spine surgeons were 77.4% and 76.1%, respectively. We demonstrated an accurate, automated detection and diagnosis of spinal schwannomas and meningiomas using the developed deep learning-based method based on MRI. This system could be valuable in supporting radiological diagnosis of spinal schwannomas and meningioma, with a potential of reducing the radiologist’s overall workload.

DOI： 10.3390/jcm12155075

Artificial intelligence in a prediction model for postendoscopic retrograde cholangiopancreatography pancreatitis

Takahashi Hidekazu, Ohno Eizaburo, Furukawa Taiki, Yamao Kentaro, Ishikawa Takuya, Mizutani Yasuyuki, Iida Tadashi, Shiratori Yoshimune, Oyama Shintaro, Koyama Junji, Mori Kensaku, Hayashi Yuichiro, Oda Masahiro, Suzuki Takahisa, Kawashima Hiroki

DIGESTIVE ENDOSCOPY 2023.7

Language：English Publisher：Digestive Endoscopy

Objectives: In this study we aimed to develop an artificial intelligence-based model for predicting postendoscopic retrograde cholangiopancreatography (ERCP) pancreatitis (PEP). Methods: We retrospectively reviewed ERCP patients at Nagoya University Hospital (NUH) and Toyota Memorial Hospital (TMH). We constructed two prediction models, a random forest (RF), one of the machine-learning algorithms, and a logistic regression (LR) model. First, we selected features of each model from 40 possible features. Then the models were trained and validated using three fold cross-validation in the NUH cohort and tested in the TMH cohort. The area under the receiver operating characteristic curve (AUROC) was used to assess model performance. Finally, using the output parameters of the RF model, we classified the patients into low-, medium-, and high-risk groups. Results: A total of 615 patients at NUH and 544 patients at TMH were enrolled. Ten features were selected for the RF model, including albumin, creatinine, biliary tract cancer, pancreatic cancer, bile duct stone, total procedure time, pancreatic duct injection, pancreatic guidewire-assisted technique without a pancreatic stent, intraductal ultrasonography, and bile duct biopsy. In the three fold cross-validation, the RF model showed better predictive ability than the LR model (AUROC 0.821 vs. 0.660). In the test, the RF model also showed better performance (AUROC 0.770 vs. 0.663, P = 0.002). Based on the RF model, we classified the patients according to the incidence of PEP (2.9%, 10.0%, and 23.9%). Conclusion: We developed an RF model. Machine-learning algorithms could be powerful tools to develop accurate prediction models.

DOI： 10.1111/den.14622

Gaussian affinity and GIoU-based loss for perforation detection and localization from colonoscopy videos

肺マイクロCT像中の小葉間隔壁抽出のための教師データ生成に関する検討

深井大輔, 小田紘久, 林雄一郎, 鄭通, 中村彰太, 小田昌宏, 森健策

日本医用画像工学会大会予稿集 Vol. 42回 page： 153 - 154 2023.7

Language：Japanese Publisher：(一社)日本医用画像工学会

本稿では,肺マイクロCT像における小葉間隔壁抽出のための教師データ生成手法について述べる.マイクロCTとよばれる高解像度CT撮像装置により,肺の微細構造を撮像可能となった.肺マイクロCT像による三次元的な肺微細構造解析に基づく病態や生体機能のさらなる解明が期待される.肺微細構造の一つである小葉間隔壁の構造を明らかにするため,我々はマイクロCT像からの小葉間隔壁の構造抽出を目指す.マイクロCT像における小葉間隔壁の教師データ作成の負担は大きい.本稿では,教師データ生成の負担軽減のため,二次元画像のみのラベル付与から,三次元画像のラベル画像を得る方法を検討する.本手法で得たラベル画像から5枚の断面画像をランダムに取り出して評価したところ,平均のDice係数は0.75だった.本手法によって得られたラベル画像を元に,手動で断面画像の教師データ生成を行ったところ,作業時間が27～30分から5～7分に短縮された.(著者抄録)

researchmap

2.5次元特徴抽出及び3次元特徴選択によるCOVID-19自動分類モデルの汎化性能に関する検討

遠田涼, 伊東隼人, 小田昌宏, 林雄一郎, 大竹義人, 橋本正弘, 明石敏昭, 青木茂樹, 森健策

日本医用画像工学会大会予稿集 Vol. 42回 page： 209 - 210 2023.7

Language：Japanese Publisher：(一社)日本医用画像工学会

本稿では,2.5次元の特徴抽出及び3次元の特徴選択によるCOVID-19自動分類用CNNを提案し,その汎化性能を複数のデータセットを用いて検証する.新型コロナウイルス感染症(COVID-19)の診断において,胸部CT検査は不可欠であるが,CT画像の読影は放射線科医らにとって負担が大きく,コンピュータ支援診断技術の開発が求められている.COVID-19の流行初期から様々な自動分類モデルが提案されてきたが,その多くは小規模な内部データセットで分類性能を検証するに留まっており,汎用的なモデルは少数である.本研究では,2次元及び3次元の特徴抽出器を組み合わせることで過学習を抑制した2.5次元のCNNと,抽出された特徴から重要なパターンのみを選択する3次元のattention機構を有する分類モデルを提案し,その汎化性能を検証する.提案モデルを含めて6種類のモデルを用いた比較の結果,提案モデルは複数のデータセットで安定した分類性能を示した.(著者抄録)

researchmap

DEVELOPMENT OF A MACHINE-LEARNING MODEL FOR PREDICTING POST-ERCP PANCREATITIS

Takahashi Hidekazu, Eizaburo Ohno, Taiki Furukawa, Kentaro Yamao, Takuya Ishikawa, Yasuyuki Mizutani, Tadashi Iida, Yoshimune Shiratori, Shintaro Oyama, Junji Koyama, Kensaku Mori, Yuichiro Hayashi, Masahiro Oda, Takahisa Suzuki, Hiroki Kawashima

Gastrointestinal Endoscopy Vol. 97 ( 6 ) page： AB656 - AB656 2023.6

Jiang Kai, Itoh Hayato, Oda Masahiro, Okumura Taishi, Mori Yuichi, Misawa Masashi, Hayashi Takemasa, Kudo Shin-Ei, Mori Kensaku

INTERNATIONAL JOURNAL OF COMPUTER ASSISTED RADIOLOGY AND SURGERY Vol. 18 ( 5 ) page： 795 - 805 2023.5

Language：English Publisher：International Journal of Computer Assisted Radiology and Surgery

Purpose: Endoscopic submucosal dissection (ESD) is a minimally invasive treatment for early gastric cancer. However, perforations may happen and cause peritonitis during ESD. Thus, there is a potential demand for a computer-aided diagnosis system to support physicians in ESD. This paper presents a method to detect and localize perforations from colonoscopy videos to avoid perforation ignoring or enlarging by ESD physicians. Method: We proposed a training method for YOLOv3 by using GIoU and Gaussian affinity losses for perforation detection and localization in colonoscopic images. In this method, the object functional contains the generalized intersection over Union loss and Gaussian affinity loss. We propose a training method for the architecture of YOLOv3 with the presented loss functional to detect and localize perforations precisely. Results: To qualitatively and quantitatively evaluate the presented method, we created a dataset from 49 ESD videos. The results of the presented method on our dataset revealed a state-of-the-art performance of perforation detection and localization, which achieved 0.881 accuracy, 0.869 AUC, and 0.879 mean average precision. Furthermore, the presented method is able to detect a newly appeared perforation in 0.1 s. Conclusions: The experimental results demonstrated that YOLOv3 trained by the presented loss functional were very effective in perforation detection and localization. The presented method can quickly and precisely remind physicians of perforation happening in ESD. We believe a future CAD system can be constructed for clinical applications with the proposed method.

DOI： 10.1007/s11548-022-02821-x