Learning to segment subcortical structures from noisy annotations with a novel uncertainty-reliability aware learning framework

CSCE: Cross Supervising and Confidence Enhancement pseudo-labels for semi-supervised subcortical brain structure segmentation.

Partly due to the use of exhaustive-annotated data, deep networks have achieved impressive performance on medical image segmentation. Medical imaging data paired with noisy annotation are, however, ubiquitous, but little is known about the effect of noisy annotation on deep learning based medical image segmentation. We studied the effect of noisy annotation in the context of mandible segmentation from CT images. First, 202 images of head and neck cancer patients were collected from our clinical database, where the organs-at-risk were annotated by one of twelve planning dosimetrists. The mandibles were roughly annotated as the planning avoiding structure. Then, mandible labels were checked and corrected by a head and neck specialist to get the reference standard. At last, by varying the ratios of noisy labels in the training set, deep networks were trained and tested for mandible segmentation. The trained models were further tested on other two public datasets. Experimental results indicated that the network trained with noisy labels had worse segmentation than that trained with reference standard, and in general, fewer noisy labels led to better performance. When using 20% or less noisy cases for training, no significant difference was found on the segmentation results between the models trained by noisy or reference annotation. Cross-dataset validation results verified that the models trained with noisy data achieved competitive performance to that trained with reference standard. This study suggests that the involved network is robust to noisy annotation to some extent in mandible segmentation from CT images. It also highlights the importance of labeling quality in deep learning. In the future work, extra attention should be paid to how to utilize a small number of reference standard samples to improve the performance of deep learning with noisy annotation.

MSGSE-Net: Multi-scale guided squeeze-and-excitation network for subcortical brain structure segmentation

Deep convolutional neural networks have driven substantial advancements in the automatic understanding of images. Requiring a large collection of images and their associated annotations is one of the main bottlenecks limiting the adoption of deep networks. In the task of medical image segmentation, requiring pixel-level semantic annotations performed by human experts exacerbate this difficulty. This paper proposes a new framework to train a fully convolutional segmentation network from a large set of cheap unreliable annotations and a small set of expert-level clean annotations. We propose a spatially adaptive reweighting approach to treat clean and noisy pixel-level annotations commensurately in the loss function. We deploy a meta-learning approach to assign higher importance to pixels whose loss gradient direction is closer to those of clean data. Our experiments on training the network using segmentation ground truth corrupted with different levels of annotation noise show how spatial reweighting improves the robustness of deep networks to noisy annotations.

Embracing imperfect datasets: A review of deep learning solutions for medical image segmentation.

In this paper, we propose an approach for super-resolution land cover mapping on remote sensing images based on a Convolutional Neural Network (CNN). Here, the CNN is trained to match the input subimages to the super resolution map around the training pixels. Since there are so many possible configurations of super-resolution map on a given set of pixels, a large number of training samples are required. To reduce the number of training samples, we converted the super-resolution to a set of level set functions and used the minimum mean square error between the predicted and actual level set functions as the training objective. The QUICKBIRD satellite image data cover a part of Kasetsart University’s Bangkhen campus was used for evaluation. Experimental results showed that the proposed method has achieved superior accuracy than both Hopfield and Pixel-Swapping methods.

Ischaemic stroke is a medical condition caused by occlusion of blood supply to the brain tissue thus forming a lesion. A lesion is zoned into a core associated with irreversible necrosis typically located at the center of the lesion, while reversible hypoxic changes in the outer regions of the lesion are termed as the penumbra. Early estimation of core and penumbra in ischaemic stroke is crucial for timely intervention with thrombolytic therapy to reverse the damage and restore normalcy. Multisequence magnetic resonance imaging (MRI) is commonly employed for clinical diagnosis. However, a sequence singly has not been found to be sufficiently able to differentiate between core and penumbra, while a combination of sequences is required to determine the extent of the damage. The challenge, however, is that with an increase in the number of sequences, it cognitively taxes the clinician to discover symptomatic biomarkers in these images. In this paper, we present a data-driven fully automated method for estimation of core and penumbra in ischaemic lesions using diffusion-weighted imaging (DWI) and perfusion-weighted imaging (PWI) sequence maps of MRI. The method employs recent developments in convolutional neural networks (CNN) for semantic segmentation in medical images. In the absence of availability of a large amount of labeled data, the CNN is trained using an adversarial approach employing cross-entropy as a segmentation loss along with losses aggregated from three discriminators of which two employ relativistic visual Turing test. This method is experimentally validated on the ISLES-2015 dataset through three-fold cross-validation to obtain with an average Dice score of 0.82 and 0.73 for segmentation of penumbra and core respectively.

Background Ultrasound image segmentation is challenging due to the low signal-to-noise ratio and poor quality of ultrasound images. With deep learning advancements, convolutional neural networks (CNNs) have been widely used for ultrasound image segmentation. However, due to the intrinsic locality of convolutional operations and the varying shapes of segmentation objects, segmentation methods based on CNNs still face challenges with accuracy and generalization. In addition, Transformer is a network architecture with self-attention mechanisms that performs well in the field of computer vision. Based on the characteristics of Transformer and CNNs, we propose a hybrid architecture based on Transformer and U-Net with joint loss for ultrasound image segmentation, referred to as TU-Net. Methods TU-Net is based on the encoder-decoder architecture and includes encoder, parallel attention mechanism and decoder modules. The encoder module is responsible for reducing dimensions and capturing different levels of feature information from ultrasound images; the parallel attention mechanism is responsible for capturing global and multiscale local feature information; and the decoder module is responsible for gradually recovering dimensions and delineating the boundaries of the segmentation target. Additionally, we adopt joint loss to optimize learning and improve segmentation accuracy. We use experiments on datasets of two types of ultrasound images to verify the proposed architecture. We use the Dice scores, precision, recall, Hausdorff distance (HD) and average symmetric surface distance (ASD) as evaluation metrics for segmentation performance. Results For the brachia plexus and fetal head ultrasound image datasets, TU-Net achieves mean Dice scores of 79.59% and 97.94%; precisions of 81.25% and 98.18%; recalls of 80.19% and 97.72%; HDs (mm) of 12.44 and 6.93; and ASDs (mm) of 4.29 and 2.97, respectively. Compared with those of the other six segmentation algorithms, the mean values of TU-Net increased by approximately 3.41%, 2.62%, 3.74%, 36.40% and 31.96% for the Dice score, precision, recall, HD and ASD, respectively.

Automatic segmentation of CT images has recently been applied in several clinical liver applications. Convolutional Neural Networks (CNNs) have shown their effectiveness in medical image segmentation in general and also in liver segmentation. However, liver image quality may vary between medical centers due to differences in the use of CT scanners, protocols, radiation dose, and contrast enhancement. In this paper, we investigate three wells known CNNs, FCN-CRF, DRIU, and V-net, for liver segmentation using data from several medical centers. We perform qualitative evaluation of the CNNs based on Dice score, Hausdorff distance, mean surface distance and false positive rate. The results show that all three CNNs achieved a mean Dice score of over 90% in liver segmentation with typical contrast enhanced CT images of the liver. p-values from paired T-test on Dice score of the three networks using Mayo dataset are larger than 0.05 suggesting that no statistical significant difference in their performance. DRIU performs the best in term of processing time. The results also demonstrate that those CNNs have reduced performance in liver segmentation in the case of low-dose and non-contrast enhanced CT images. In conclusion, these promising results enable further investigation of alternative deep learning based approaches to liver segmentation using CT images.

With the development of computer vision and image segmentation technology, medical image segmentation and recognition technology has become an important part of computer-aided diagnosis. The traditional image segmentation method relies on artificial means to extract and select information such as edges, colors, and textures in the image. It not only consumes considerable energy resources and people's time but also requires certain expertise to obtain useful feature information, which no longer meets the practical application requirements of medical image segmentation and recognition. As an efficient image segmentation method, convolutional neural networks (CNNs) have been widely promoted and applied in the field of medical image segmentation. However, CNNs that rely on simple feedforward methods have not met the actual needs of the rapid development of the medical field. Thus, this paper is inspired by the feedback mechanism of the human visual cortex, and an effective feedback mechanism calculation model and operation framework is proposed, and the feedback optimization problem is presented. A new feedback convolutional neural network algorithm based on neuron screening and neuron visual information recovery is constructed. So, a medical image segmentation algorithm based on a feedback mechanism convolutional neural network is proposed. The basic idea is as follows: The model for obtaining an initial region with the segmented medical image classifies the pixel block samples in the segmented image. Then, the initial results are optimized by threshold segmentation and morphological methods to obtain accurate medical image segmentation results. Experiments show that the proposed segmentation method has not only high segmentation accuracy but also extremely high adaptive segmentation ability for various medical images. The research in this paper provides a new perspective for medical image segmentation research. It is a new attempt to explore more advanced intelligent medical image segmentation methods. It also provides technical approaches and methods for further development and improvement of adaptive medical image segmentation technology.

Accurate segmentation of medical images is vital for disease detection and treatment. Convolutional Neural Networks (CNN) and Transformer models are widely used in medical image segmentation due to their exceptional capabilities in image recognition and segmentation. However, CNNs often lack an understanding of the global context and may lose spatial details of the target, while Transformers struggle with local information processing, leading to reduced geometric detail of the target. To address these issues, this research presents a Global-Local Fusion network model (GLFUnet) based on the U-Net framework and attention mechanisms. The model employs a dual-branch network that utilizes ConvNeXt and Swin Transformer to simultaneously extract multi-level features from pathological images. It enhances ConvNeXt’s local feature extraction with spatial and global attention up-sampling modules, while improving Swin Transformer’s global context dependency with channel attention. The Attention Feature Fusion module and skip connections efficiently merge local detailed and global coarse features from CNN and Transformer branches at various scales. The fused features are then progressively restored to the original image resolution for pixel-level prediction. Comprehensive experiments on datasets of stomach and liver cancer demonstrate GLFUnet’s superior performance and adaptability in medical image segmentation, holding promise for clinical analysis and disease diagnosis.

Cascaded deep convolutional encoder-decoder neural networks for efficient liver tumor segmentation

Deformable image registration is a fundamental task in medical imaging. Due to the large computational complexity of deformable registration of volumetric images, conventional iterative methods usually face the tradeoff between the registration accuracy and the computation time in practice. In order to boost the performance of deformable registration in both accuracy and runtime, we propose a fast unsupervised convolutional neural network for deformable image registration. The proposed registration model FDRN possesses a compact encoder-decoder network architecture which employs a pair of fixed and moving images as input and outputs a three-dimensional displacement vector field (DVF) describing the offsets between the corresponding voxels in the fixed and moving images. In order to efficiently utilize the memory resources and enlarge the model capacity, we adopt additive forwarding instead of channel concatenation and deepen the network in each encoder and decoder stage. To facilitate the learning efficiency, we leverage skip connection within the encoder and decoder stages to enable residual learning and employ an auxiliary loss at the bottom layer with lowest resolution to involve deep supervision. Particularly, the low-resolution auxiliary loss is weighted by an exponentially decayed parameter during the training phase. In conjunction with the main loss in high-resolution grid, a coarse-to-fine learning strategy is achieved. Last but not least, we involve a proposed multi-label segmentation loss (SL) to improve the network performance in Dice score in case the segmentation prior is available. Comparing to the SL using average Dice score, the proposed SL does not require additional memory in the training phase and improves the registration accuracy efficiently. We evaluated FDRN on multiple brain MRI datasets from different aspects including registration accuracy, model generalizability, and model analysis. Experimental results demonstrate that FDRN performs better than the state-of-the-art registration method VoxelMorph by 1.46% in Dice score in LPBA40. In addition to LPBA40, FDRN obtains the best Dice and NCC among all the investigated methods in the unseen MRI datasets including CUMC12, MGH10, ABIDE, and ADNI by a large margin. The proposed FDRN provides better performance than the existing state-of-the-art registration methods for brain MR images by resorting to the compact autoencoder structure and efficient learning. Additionally, FDRN is a generalized framework for image registration which is not confined to a particular type of medical images or anatomy.

Medical image segmentation algorithm based on feedback mechanism convolutional neural network

Medical image segmentation is crucial for improving healthcare outcomes. Convolutional neural networks (CNNs) have been widely applied in medical image analysis; however, their inherent inductive biases limit their ability to capture global contextual information. Vision transformer (ViT) architectures address this limitation by leveraging attention mechanisms to model global relationships; however, they typically require large-scale datasets for effective training, which is challenging in the field of medical imaging due to limited data availability. This study aimed to integrate the advantages of CNN and ViT architectures to improve segmentation performance on small-scale medical image datasets. In this study, we established a U-shaped network architecture based on a Transformer-assisted convolutional neural network (TAC-UNet). The TAC-UNet is primarily composed of a hybrid structure integrating CNN and Transformer components. Specifically, the hybrid architecture follows a dual-path design in which the Transformer branch continuously conveys global contextual information to the CNN backbone. This allows the CNN backbone to enhance its global perception while building on the local features it extracts, thereby improving its ability to comprehend complex image structures. A channel cross-attention (CCA) module is also incorporated as a bridge between the encoder and decoder to better reconcile the semantic discrepancies between them. Detailed experiments on three public datasets were conducted. Specifically, our model was trained on 30 images from the Multi-organ Nucleus Segmentation (MoNuSeg) training dataset, 85 images from the Gland Segmentation (GlaS) training dataset, and 551 images from the Computer Vision Center Colorectal Cancer-Clinic Database (CVC-ClinicDB) dataset. We evaluated the performance of our model on the corresponding test sets. Our TAC-UNet achieved the best Dice scores (80.36%, 90.70%, and 91.81% on the MoNuSeg, GlaS, and CVC-ClinicDB datasets, respectively) of all the models. Compared to other CNN-based, Transformer-based, and hybrid methods, the TAC-UNet demonstrated significantly superior segmentation performance. Our TAC-UNet model showed advanced segmentation performance on small-scale medical image datasets. The detailed experimental results showed the effectiveness of the method. Our model's code is available at: https://github.com/hejlhello/TAC-UNet.