Medical image segmentation methods downsample images for feature extraction and then upsample them to restore resolution for pixel-level predictions. In such schema, upsample technique is vital in restoring informatio...
详细信息
ISBN:
(数字)9781665468190
ISBN:
(纸本)9781665468206
Medical image segmentation methods downsample images for feature extraction and then upsample them to restore resolution for pixel-level predictions. In such schema, upsample technique is vital in restoring information for better performance. However, existing upsample techniques leverage little information from downsampling paths. The local and detailed feature from the shallower layer such as boundary and tissue texture is crucial in segmentation, especially medical image segmentation. To this end, we propose a novel upsample approach for medical image segmentation, Window Attention Upsample (WAU), which upsamples features conditioned on local and detailed features from downsampling path in local windows by introducing attention decoders of Transformer. WAU could serve as a general upsample method and be incorporated into any segmentation model that possesses lateral connections. We first propose the Attention Upsample which consists of Attention Decoder (AD) and bilinear upsample. AD leverages pixel-level attention to model longrange dependency and global information for a better upsample. Bilinear upsample is introduced as the residual connection to complement the upsampled features. Moreover, considering the extensive memory and computation cost of pixel-level attention, we further design a window attention scheme to restrict attention computation in local windows instead of the global range. We evaluate our method (WAU) on classic UNet structure with lateral connections and achieve state-of-the-art performance on Medical Segmentation Decathlon (MSD) Brain and Automatic Cardiac Diagnosis Challenge (ACDC) datasets. We also validate the effectiveness of our method on multiple classic architectures and achieve consistent improvement.
Histopathology image segmentation is crucial in disease diagnosis, therapeutic response evaluation, and prognosis. However, manually annotating pixel-level labels for histopathology images is both time-consuming and l...
Histopathology image segmentation is crucial in disease diagnosis, therapeutic response evaluation, and prognosis. However, manually annotating pixel-level labels for histopathology images is both time-consuming and labor-demanding task. In this study, we propose a novel semi-supervised semantic segmentation framework called UTCS ( U ncertainty-guided cross T eaching and C urriculum S elf-training) to address the challenges of limited labeled data. UTCS effectively harnesses the benefits of consistency regularization and self-training in semi-supervised learning. Our approach introduces a mutual consistency network, where one network’s prediction is used as a pseudo mask to supervise the other network and vice versa. Addressing the issue of unreliable pseudo labels, we propose a dynamically re-weighted loss function that leverages uncertainty to perform pixel-level selection during the mutual teaching process, referred to as uncertainty-guided cross teaching. Furthermore, inspiring from curriculum learning, we incorporate an self-training strategy, focusing on image-level selection, that prioritizes reliable images during the re-training stage and aims to generate high-quality pseudo-labels for less reliable images. Extensive experiments on two publicly available histopathology datasets, BCSS and LUAD-HistoSeg, demonstrate the superior performance of our method compared to state-of-the-art semi-supervised semantic segmentation methods.
Moving object detection is an important application of computer vision. Commonly used foreground separation algorithms such as Gaussian mixture modeling, ViBe, frame difference method, etc., do not consider the color ...
详细信息
ISBN:
(数字)9781728152448
ISBN:
(纸本)9781728152455
Moving object detection is an important application of computer vision. Commonly used foreground separation algorithms such as Gaussian mixture modeling, ViBe, frame difference method, etc., do not consider the color of shadow and recognize the shadow of a moving object as a part of the moving object. In many cases, the shadow detection effect is not good. Focus on the detection of moving object shadows in traffic surveillance videos, this paper improves the existing ViBe algorithm, considers the color characteristics of the shadows, recognizes the shadows as part of the background, gains a smaller amount of calculation and better effect of shadow detection with the advantages of ViBe.
Medical image segmentation methods downsample images for feature extraction and then upsample them to restore resolution for pixel-level predictions. In such schema, upsample technique is vital in restoring informatio...
详细信息
The development of an Artificial Intelligence (AI) assisted tissue segmentation method of digital pathology images is critical for cancer diagnosis and prognosis. Excellent performance has been achieved with the curre...
详细信息
The main appearance difference between different types of vehicles is located in the front face area, so the car face parts detection is a key role in fine-grained vehicle recognition. This paper presents a faster R-C...
详细信息
Sparse subspace learning has been demonstrated to be effective in data mining and machine learning. In this paper, we cast the unsupervised feature selection scenario as a matrix factorization problem from the viewpoi...
详细信息
Sparse subspace learning has been demonstrated to be effective in data mining and machine learning. In this paper, we cast the unsupervised feature selection scenario as a matrix factorization problem from the viewpoint of sparse subspace learning. By minimizing the reconstruction residual, the learned feature weight matrix with the l 2,1 -norm and the non-negative constraints not only removes the irrelevant features, but also captures the underlying low dimensional structure of the data points. Meanwhile in order to enhance the model's robustness, l 1 -norm error function is used to resistant to outliers and sparse noise. An efficient iterative algorithm is introduced to optimize this non-convex and non-smooth objective function and the proof of its convergence is given. Although, there is a subtraction item in our multiplicative update rule, we validate its non-negativity. The superiority of our model is demonstrated by comparative experiments on various original datasets with and without malicious pollution.
Although it has become an accepted lay view that when labeling objects through crowdsourcing systems, non-expert annotators often exhibit biases, this argument lacks sufficient evidential observation and systematic em...
Although it has become an accepted lay view that when labeling objects through crowdsourcing systems, non-expert annotators often exhibit biases, this argument lacks sufficient evidential observation and systematic empirical study. This paper initially analyzes eight real-world datasets from different domains whose class labels were collected from crowdsourcing systems. Our analyses show that biased labeling is a systematic tendency for binary categorization; in other words, for a large number of annotators, their labeling qualities on the negative class (supposed to be the majority) are significantly greater than are those on the positive class (minority). Therefore, the paper empirically studies the performance of four existing EM-based consensus algorithms , DS, GLAD, RY, and ZenCrowd, on these datasets. Our investigation shows that all of these state-of-the-art algorithms ignore the potential bias characteristics of datasets and perform badly although they model the complexity of the systems. To address the issue of handling biased labeling, the paper further proposes a novel consensus algorithm, namely adaptive weighted majority voting (AWMV), based on the statistical difference between the labeling qualities of the two classes. AWMV utilizes the frequency of positive labels in the multiple noisy label set of each example to obtain a bias rate and then assigns weights derived from the bias rate to negative and positive labels. Comparison results among the five consensus algorithms (AWMV and the four existing) show that the proposed AWMV algorithm has the best overall performance. Finally, this paper notes some potential related topics for future study.
In order to reduce traffic accidents caused by the pedestrian, five kinds of dangerous pedestrian abnormal behaviors are studied in the paper. A behavior model between the pedestrian trajectory and the road is built t...
详细信息
暂无评论