Attentive neural cell instance segmentation

Yi, Jingru; Wu, Pengxiang; Jiang, Menglin; Huang, Qi‐Fang; Hoeppner, Daniel J.; Metaxas, Dimitris N.

doi:10.1016/j.media.2019.05.004

Cited by 79 publications

(55 citation statements)

References 19 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…We use the AP at mask-level IOU [3,6] at threshold of 0.5 and 0.7 to evaluate the instance segmentation performances. We also report the mean mask-level IOU [12] between the predicted segmentation masks and the ground truth masks at threshold of 0.5 and 0.7. Table 1.…”

Section: Methodsmentioning

confidence: 99%

Multi-scale Cell Instance Segmentation with Keypoint Graph Based Bounding Boxes

Huang

et al. 2019

Lecture Notes in Computer Science

Self Cite

View full text Add to dashboard Cite

Most existing methods handle cell instance segmentation problems directly without relying on additional detection boxes. These methods generally fails to separate touching cells due to the lack of global understanding of the objects. In contrast, box-based instance segmentation solves this problem by combining object detection with segmentation. However, existing methods typically utilize anchor boxbased detectors, which would lead to inferior instance segmentation performance due to the class imbalance issue. In this paper, we propose a new box-based cell instance segmentation method. In particular, we first detect the five pre-defined points of a cell via keypoints detection. Then we group these points according to a keypoint graph and subsequently extract the bounding box for each cell. Finally, cell segmentation is performed on feature maps within the bounding boxes. We validate our method on two cell datasets with distinct object shapes, and empirically demonstrate the superiority of our method compared to other instance segmentation techniques. Code is available at: https: //github.com/yijingru/KG_Instance_Segmentation.

show abstract

Section: Methodsmentioning

confidence: 99%

Multi-scale Cell Instance Segmentation with Keypoint Graph Based Bounding Boxes

Huang

et al. 2019

Lecture Notes in Computer Science

Self Cite

View full text Add to dashboard Cite

show abstract

“…Object detection networks, such as Faster R-CNN [24], Single Shot Detectors (SSD) [25] and You Only Look Once (YOLO) [26], have been implemented in a number of instance segmentation neural networks in addition to Mask R-CNN. Attentive neural cell instance segmentation (ANCIS) applies UNet based segmentation on regions proposed by an SSD [14]. In addition, ANCIS applies two small CNN modules to locate cell features and suppress unwanted noise regions, potentially enhancing accurate delineation of intricate structures and boarders.…”

Section: Region Proposal Based Segmentation Networkmentioning

confidence: 99%

“…All networks were implemented using Python 3.6.9 with Tensor ow 1. 14 All networks utilized the Adam optimization algorithm to iteratively update network weights. Additionally, all networks used image augmentation to increase the robustness of the training datasets.…”

Section: B Neural Network Implementationmentioning

confidence: 99%

Application of Convolutional Neural Networks Towards Nuclei Segmentation in Localization-based Super-Resolution Fluorescence Microscopy Images

Mela

Liu

2021

Preprint

View full text Add to dashboard Cite

Background Automated segmentation of nuclei in microscopic images has been conducted to enhance throughput in pathological diagnostics and biological research. Segmentation accuracy and speed has been significantly enhanced with the advent of convolutional neural networks. A barrier in the broad application of neural networks to nuclei segmentation is the necessity to train the network using a set of application specific images and image labels. Previous works have attempted to create broadly trained networks for universal nuclei segmentation; however, such networks do not work on all imaging modalities, and best results are still commonly found when the network is retrained on user specific data. Stochastic optical reconstruction microscopy (STORM) based super-resolution fluorescence microscopy has opened a new avenue to image nuclear architecture at nanoscale resolutions. Due to the large size and discontinuous features typical of super-resolution images, automatic nuclei segmentation can be difficult. In this study, we apply commonly used networks (Mask R-CNN and UNet architectures) towards the task of segmenting super-resolution images of nuclei. First, we assess whether networks broadly trained on conventional fluorescence microscopy datasets can accurately segment super-resolution images. Then, we compare the resultant segmentations with results obtained using networks trained directly on our super-resolution data. We next attempt to optimize and compare segmentation accuracy using three different neural network architectures. Results Results indicate that super-resolution images are not broadly compatible with neural networks trained on conventional bright-field or fluorescence microscopy images. When the networks were trained on super-resolution data, however, we attained nuclei segmentation accuracies (F1-Score) in excess of 0.8, comparable to past results found when conducting nuclei segmentation on conventional fluorescence microscopy images. Overall, we achieved the best results utilizing the Mask R-CNN architecture. Conclusions We found that convolutional neural networks are powerful tools capable of accurately and quickly segmenting localization-based super-resolution microscopy images of nuclei. While broadly trained and widely applicable segmentation algorithms are desirable for quick use with minimal input, optimal results are still found when the network is both trained and tested on visually similar images. We provide a set of Colab notebooks to disseminate the software into the broad scientific community (https://github.com/YangLiuLab/Super-Resolution-Nuclei-Segmentation).

show abstract

“…Hua zhu Fu, et al [36] proposed the M-Net that mainly consists of multi-scale input layer, U-shape convolutional network, side-output layer, and multi-label loss function. Jingru Yi, et al [37] proposed an attentive cell instance segmentation method that builds on a joint network that combines a U-net and a single shot multi-box detector. Nisha Ramesh, et al [38] uses multi-task learning in combination with the similarity interface to detect and segment cells in microscopy images.…”

Section: Previous Related Workmentioning

confidence: 99%

Segmentation of Cell Images Based on Improved Deep Learning Approach

2020

View full text Add to dashboard Cite

The improved U_net algorithm based on mixed convolution blocks (McbUnet), which combines the advantages of U-Net and residual learning, is proposed for cell image segmentation in this paper. The network is mainly composed of two kinds of mixed convolution blocks. There are three main benefits to this algorithm. First, the convolution block can utilize different size kernels to overcome the limitation of a single size convolution kernel in traditional deep convolution. Second, in the mixed convolution blocks, two hyperparameters (Width multiplier and Resolution multiplier) are used to quickly adjust the model to fit a specific environment. Third, the residual paths are improved. We test the proposed network and compare it with other recent segmentation methods based on deep learning. The proposed method is superior to comparison methods, which shows its effectiveness. INDEX TERMS microscopy images; deep learning; cell segmentation; convolutional neural network

show abstract

Attentive neural cell instance segmentation

Cited by 79 publications

References 19 publications

Multi-scale Cell Instance Segmentation with Keypoint Graph Based Bounding Boxes

Multi-scale Cell Instance Segmentation with Keypoint Graph Based Bounding Boxes

Application of Convolutional Neural Networks Towards Nuclei Segmentation in Localization-based Super-Resolution Fluorescence Microscopy Images

Segmentation of Cell Images Based on Improved Deep Learning Approach

Contact Info

Product

Resources

About