Abstract
BackgroundWith the development of deep learning (DL), more and more methods based on deep learning are proposed and achieve state-of-the-art performance in biomedical image segmentation. However, these methods are usually complex and require the support of powerful computing resources. According to the actual situation, it is impractical that we use huge computing resources in clinical situations. Thus, it is significant to develop accurate DL based biomedical image segmentation methods which depend on resources-constraint computing.ResultsA lightweight and multiscale network called PyConvU-Net is proposed to potentially work with low-resources computing. Through strictly controlled experiments, PyConvU-Net predictions have a good performance on three biomedical image segmentation tasks with the fewest parameters.ConclusionsOur experimental results preliminarily demonstrate the potential of proposed PyConvU-Net in biomedical image segmentation with resources-constraint computing.
Highlights
With the development of deep learning (DL), more and more methods based on deep learning are proposed and achieve state-of-the-art performance in biomedical image segmentation
The last dataset is ISBICell [22] is provided by the EM segmentation challenge that was started at ISBI 2012 and is still open for new contributions
In our studies, we observe that U-Net will ignore detailed information when performing convolution operations [27]. We analyze this issue in detail and address it by proposing a lightweight and multiscale architecture PyConvU-Net which replaces the traditional convolution layer with the pyramidal convolution layer
Summary
With the development of deep learning (DL), more and more methods based on deep learning are proposed and achieve state-of-the-art performance in biomedical image segmentation. These methods are usually complex and require the support of powerful computing resources. It is significant to develop accurate DL based biomedical image segmentation methods which depend on resources-constraint computing. Results: A lightweight and multiscale network called PyConvU-Net is proposed to potentially work with low-resources computing. Conclusions: Our experimental results preliminarily demonstrate the potential of proposed PyConvU-Net in biomedical image segmentation with resources-constraint computing. With the rapid developments of DL based techniques, multiple researchers begin to investigate the potential applications to employ DL in biomedical image segmentation. Since the U-Net architecture was proposed in 2015, more and more researchers choose it as the backbone for their models
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
