Search

• A Polarmask-based method for blood cell contour segmentation is proposed, which uses weak labels to train the model to obtain a pre-training weight. Using this pre-training weight not only improves the training speed of the model but also improves the model's accuracy for cell image instance segmentation. • We designed the smoothing constraint loss based on the shape properties of cells in blood cell images, which makes the segmented cell contours smoother. • A spatial attention mechanism is added to the backbone network, effectively improving model segmentation accuracy. • Our method can help healthcare workers quickly identify the number of cells and cell shapes, which reduces the workload for healthcare workers and has good medical value. A PolarMask-based method for blood cell contour segmentation is proposed. The method is divided into two parts. One part is a weak label-based model pretraining method, which uses weak labels to train the model and obtain a pretraining weight. The training speed and accuracy of the segmentation model are accelerated. The other part is based on the PolarMask method to segment the white and red blood cells in blood cells and can obtain smoother cell contours. Thus, it improves the accuracy of blood cell segmentation. Our method can help medical personnel identify the number of cells and cell shape quickly, which reduces the workload for medical personnel. We improve PolarMask to make it more suitable for blood cell contour segmentation, and the improved method can be divided into two parts. In the first part, we use a weakly labeled dataset with the labeling type of bounding boxes for pretraining and then use the labels of the segmentation type for transfer learning of the cell segmentation model. In the second part, we add a smoothing constraint loss to the loss function of the mask to smoothen the segmented cell contours. We add the SE attention mechanism in the backbone network (ResNet18) to further improve the segmentation accuracy. Our method is mainly used for the segmentation of blood cell (erythrocyte and leukocyte) contours. Our method improves average precision (AP) by 8.4% and AP 50 by 0.6% compared with PolarMask. The most significant improvement is in AP 75 , which improves by 8.8%. Our method models blood cell contours based on PolarMask and uses a weakly labeled training model to obtain pretrained weights that can segment red and white blood cells. Our method effectively improves the accuracy of the model in segmenting blood cells, and the segmented blood cell contours are smoother. [ABSTRACT FROM AUTHOR]

• This paper summarizes the important role of current computer-aided segmentation in cell segmentation. • This paper systematically reviews image segmentation methods based on traditional and deep learning methods. • This paper, from the three aspects of cell image weak label extraction, cell image instance segmentation, and cell internal structure segmentation, deep-learning-based cell image segmentation methods are analyzed and summarized. • This paper summarizes the cell image instance segmentation and discusses its challenges and future developments. The instance segmentation of cell images is the basis for conducting cell research and is of great importance for the study and diagnosis of pathologies. To analyze current situations and future developments in the field of cell image instance segmentation, this paper first systematically reviews image segmentation methods based on traditional and deep learning methods. Then, from the three aspects of cell image weak label extraction, cell image instance segmentation, and cell internal structure segmentation, deep-learning-based cell image segmentation methods are analyzed and summarized. Finally, cell image instance segmentation is summarized, and challenges and future developments are discussed. [ABSTRACT FROM AUTHOR]

Mesh is an essential and effective data representation of a 3D shape. The 3D mesh segmentation is a fundamental task in computer vision and graphics. It has recently been realized through a multi-scale deep learning framework, whose sampling methods are of key significance. Rarely do the previous sampling methods consider the receptive field contour of vertex, leading to loss in scale consistency of the vertex feature. Meanwhile, uniform sampling can ensure the utmost uniformity of the vertex distribution of the sampled mesh. Consequently, to efficiently improve the scale consistency of vertex features, uniform sampling was first used in this study to construct a multi-scale mesh hierarchy. In order to address the issue on uniform sampling, namely, the smoothing effect, vertex clustering sampling was used because it can preserve the geometric structure, especially the edge information. With the merits of these two sampling methods combined, more and complete information on the 3D shape can be acquired. Moreover, we adopted the attention mechanism to better realize the cross-scale shape feature transfer. According to the attention mechanism, shape feature transfer between different scales can be realized by the construction of a novel graph structure. On this basis, we propose dual-sampling attention pooling for graph neural networks on 3D mesh. According to experiments on three datasets, the proposed methods are highly competitive. [ABSTRACT FROM AUTHOR]

• This paper proposes a multiswarm Artificial Bee Colony multi-objective optimization algorithm (MS-ABC).The algorithm can be applied to accelerate the solution of complex problems on Spark platform. • Compared with the traditional algorithm, the algorithm greatly improves the speed of data processing, especially in the situation of large amount of data and high complexity. • The algorithm runs on the Spark platform to accelerate the solution of complex application problems. It has excellent performance in the medical image registration. Over the years, medical image registration has been widely used in various fields. However, different application characteristics, such as scale, computational complexity, and optimization goals, can cause problems. Therefore, developing an optimization algorithm based on clustering calculation is crucial. To solve the aforementioned problem, a multiswarm artificial bee colony (MS-ABC) multi-objective optimization algorithm based on clustering calculation is proposed. This algorithm can accelerate the resolution of complex problems on the Spark platform. Experiments show that the algorithm can optimize certain conventional complex problems and perform medical image registration tests. Results show that the MS-ABC algorithm demonstrates excellent performance in medical image registration tests. The optimization results of the MS-ABC algorithm for conventional problems are similar to those of existing algorithms; however, its performance is more time efficient for complex problems, especially when additional goals are needed. The MS-ABC algorithm is applied to the Spark platform to accelerate the resolution of complex application problems. It can solve the problem of traditional algorithms regarding long calculation time, especially in the case of highly complex and large amounts of data, which can substantially improve data-processing efficiency. [ABSTRACT FROM AUTHOR]