Search

Several factors are associated with the success of deep learning. One of the most important reasons is the availability of large-scale datasets with clean annotations. However, obtaining datasets with accurate labels in the medical imaging domain is challenging. The reliability and consistency of medical labeling are some of these issues, and low-quality annotations with label noise usually exist. Because noisy labels reduce the generalization performance of deep neural networks, learning with noisy labels is becoming an essential task in medical image analysis. Literature on this topic has expanded in terms of volume and scope. However, no recent surveys have collected and organized this knowledge, impeding the ability of researchers and practitioners to utilize it. In this work, we presented an up-to-date survey of label-noise learning for medical image domain. We reviewed extensive literature, illustrated some typical methods, and showed unified taxonomies in terms of methodological differences. Subsequently, we conducted the methodological comparison and demonstrated the corresponding advantages and disadvantages. Finally, we discussed new research directions based on the characteristics of medical images. Our survey aims to provide researchers and practitioners with a solid understanding of existing medical label-noise learning, such as the main algorithms developed over the past few years, which could help them investigate new methods to combat with the negative effects of label noise., Competing Interests: Declaration of competing interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper., (Copyright © 2024. Published by Elsevier B.V.)

Background: Breast cancer is the most common cancer among women, and ultrasound is a usual tool for early screening. Nowadays, deep learning technique is applied as an auxiliary tool to provide the predictive results for doctors to decide whether to make further examinations or treatments. This study aimed to develop a hybrid learning approach for breast ultrasound classification by extracting more potential features from local and multi-center ultrasound data., Methods: We proposed a hybrid learning approach to classify the breast tumors into benign and malignant. Three multi-center datasets (BUSI, BUS, OASBUD) were used to pretrain a model by federated learning, then every dataset was fine-tuned at local. The proposed model consisted of a convolutional neural network (CNN) and a graph neural network (GNN), aiming to extract features from images at a spatial level and from graphs at a geometric level. The input images are small-sized and free from pixel-level labels, and the input graphs are generated automatically in an unsupervised manner, which saves the costs of labor and memory space., Results: The classification AUC ROC of our proposed method is 0.911, 0.871 and 0.767 for BUSI, BUS and OASBUD. The balanced accuracy is 87.6%, 85.2% and 61.4% respectively. The results show that our method outperforms conventional methods., Conclusions: Our hybrid approach can learn the inter-feature among multi-center data and the intra-feature of local data. It shows potential in aiding doctors for breast tumor classification in ultrasound at an early stage., (© 2024. The Author(s).)

The annotation of brain lesion images is a key step in clinical diagnosis and treatment of a wide spectrum of brain diseases. In recent years, segmentation methods based on deep learning have gained unprecedented popularity, leveraging a large amount of data with high-quality voxel-level annotations. However, due to the limited time clinicians can provide for the cumbersome task of manual image segmentation, semi-supervised medical image segmentation methods present an alternative solution as they require only a few labeled samples for training. In this paper, we propose a novel semi-supervised segmentation framework that combines improved mean teacher and adversarial network. Specifically, our framework consists of (i) a student model and a teacher model for segmenting the target and generating the signed distance maps of object surfaces, and (ii) a discriminator network for extracting hierarchical features and distinguishing the signed distance maps of labeled and unlabeled data. Besides, based on two different adversarial learning processes, a multi-scale feature consistency loss derived from the student and teacher models is proposed, and a shape-aware embedding scheme is integrated into our framework. We evaluated the proposed method on the public brain lesion datasets from ISBI 2015, ISLES 2015, and BRATS 2018 for the multiple sclerosis lesion, ischemic stroke lesion, and brain tumor segmentation respectively. Experiments demonstrate that our method can effectively leverage unlabeled data while outperforming the supervised baseline and other state-of-the-art semi-supervised methods trained with the same labeled data. The proposed framework is suitable for joint training of limited labeled data and additional unlabeled data, which is expected to reduce the effort of obtaining annotated images., Competing Interests: Declaration of Competing Interest None., (Copyright © 2021. Published by Elsevier Inc.)

Purpose: Accurately delineating clinical target volumes (CTV) is essential for completing radiotherapy plans but is time-consuming, labor-intensive, and prone to inter-observer variation. Automating CTV delineation has the benefits of both speeding up contouring process and improving the quality of contours. Recently, auto-segmentation approaches based on deep learning have achieved some improvements. However, unlike organ segmentation, the CTV contains potential tumor spread tissues or subclinical disease tissues, resulting in poorly defined margin interface and irregular shape. It is not reasonable to directly apply the deep learning segmentation algorithms to CTV tasks without considering the unique characteristics of shape and margin. In this work, we propose a novel automatic CTV delineation algorithm based on deep learning addressing the unique shape and margin challenges., Methods: Our deep learning method, called RA-CTVNet, segments the CTV from cervical cancer CT images. RA-CTVNet denotes our automatic CTV delineation algorithm based on deep learning with Area-aware reweight strategy and Recursive refinement strategy. (1) In order to process the whole-volume CT images and delineate all CTVs in one shot, our method is built upon the popular 3D Unet architecture. We further extend it with robust residual learning and squeeze-and-excitation blocks for better feature representation. (2) We propose area-aware reweight strategy which assigns different weights for different slices. The core is adjusting model's attention to each slice. (3) In terms of the trade-off between providing performance improvements and meeting the limitations of GPU memory, we exploit a new recursive refinement strategy to address margin challenge., Results: This retrospective study included 462 patients diagnosed with cervical cancer who received radiotherapy from June 2017 to May 2019. Extensive experiments were conducted to evaluate performance of RA-CTVNet. First, compared to different network architectures, RA-CTVNet achieved improvements in Dice similarity coefficient (DSC). Second, we conducted ablation study. The results showed that compared to the backbone, area-aware reweight strategy increased DSC by 3.3% on average and recursive refinement strategy further increased DSC by 1.6% on average. Then, we compared our method with three human experts. Our RA-CTVNet performed better than two experts while comparably to the third expert. Finally, a multicenter evaluation was conducted to verify the accuracy and generalizability., Conclusions: Our findings show that deep learning is able to offer an efficient framework for automatic CTV delineation. The tailored RA-CTVNet can improve the quality of CTV contours, which has great potential for reducing the burden of experts and increasing the accuracy of delineation. In the future, if with more training data, further improvements are possible, bringing this approach closer to real clinical practice., (© 2021 American Association of Physicists in Medicine.)

• First Chinese tray-meal dataset collected (1185 images, 99 dish categories). • Method developed for dish recognition based on deep learning. • Automatic nutrition estimation method for Chinese tray meals. Tray meal is a popular way of eating in China, and tray-based automatic dietary assessment is important for public health. Relevant research is lacking because public tray meal datasets and suitable methods are unavailable. In this study, we established and published the first Chinese tray meal dataset, the ChinaLunchTray-99. We collected real-world 1185 tray meal images, covering 99 dish categories with corresponding manually annotated bounding box and category-level labels. We developed a new framework for automatic dietary assessment, which consists of dish image recognition, volume estimation and nutrition mapping. First, we demonstrated a tray meal detection model considering feature extraction, anchor scales, and loss function, resulting in a high mean Average Precision of 92.13%. Second, we proposed an automatic method to estimate volume via detection results and tray's information. Finally, nutrients were mapped from the estimated volume. Our research can promote applications of automatic dietary assessment for Chinese tray meals. [ABSTRACT FROM AUTHOR]

Deep-learning models require large amounts of accurately labeled data. However, for medical image segmentation, high-quality labels rely on expert experience, and less-experienced operators provide noisy labels. How one might mitigate the negative effects caused by noisy labels for 3D medical image segmentation has not been fully investigated. In this paper, our purpose is to propose a novel hybrid robust-learning architecture to combat noisy labels for 3D medical image segmentation. Our method consists of three components. First, we focus on the noisy annotations of slices and propose a slice-level label-quality awareness method, which automatically generates label-quality scores for slices in a set. Second, we propose a shape-awareness regularization loss based on distance transform maps to introduce prior shape information and provide extra performance gains. Third, based on a re-weighting strategy, we propose an end-to-end hybrid robust-learning architecture to weaken the negative effects caused by noisy labels. Extensive experiments are performed on two representative datasets (i.e., liver segmentation and multi-organ segmentation). Our hybrid noise-robust architecture has shown competitive performance, compared to other methods. Ablation studies also demonstrate the effectiveness of slice-level label-quality awareness and a shape-awareness regularization loss for combating noisy labels. [ABSTRACT FROM AUTHOR]