Related Experiment Video
Updated: Jan 29, 2026

Application of Deep Learning-Based Medical Image Segmentation via Orbital Computed Tomography
Published on: November 30, 2022
A New Encoding Architecture Based on Shift Multilayer Perceptron and Transformer for Medical Image Segmentation
Hepeng Zhong1, Jieqiong Yang2, Yingfei Wu1
1College of Advanced Interdisciplinary Studies, Central South University of Forestry and Technology, Changsha 410004, China.
This study introduces a new Multilayer Perceptron-Transformer (MPT) for medical image segmentation, improving accuracy in delineating cardiac and multi-organ structures. The MPT model enhances diagnostic reliability by capturing detailed anatomical features.
Area of Science:
- Medical Imaging
- Artificial Intelligence
- Computer Vision
Background:
- Accurate medical image segmentation is vital for clinical diagnosis.
- Existing methods struggle with low-level details and feature consistency, impacting reliability.
Purpose of the Study:
- To propose a novel Multilayer Perceptron-Transformer (MPT) encoding architecture.
- To improve the accuracy and reliability of medical image segmentation.
Main Methods:
- Integrated Shift Multilayer Perceptron and Transformer mechanisms.
- Utilized a SENet-based Atrous Spatial Pyramid Pooling module for multi-scale features.
- Employed a channel-feature aggregation attention module for enhanced information flow.
Main Results:
- Achieved an average Dice Similarity Coefficient (DSC) of 87.01% on the cardiac dataset.
- Obtained an average DSC of 79.35% and 95% Haus Dorff Distance of 20.07 mm on the multi-organ dataset.
- Demonstrated effective capture of local and global anatomical structures.
Conclusions:
- The proposed MPT architecture significantly enhances medical image segmentation.
- MPT provides reliable support for clinical diagnosis by improving accuracy and detail capture.
Related Concept Videos
Bacterial Transformation
Griffith made an unexpected discovery when he killed the pathogenic strain and mixed its remains with the live, non-pathogenic strain. Not only did the mixture kill host mice, but it also contained living pathogenic bacteria that...
Encoding
Automatic processing involves the encoding of details like time, space, frequency, and the meaning of words, usually done without conscious...
Inhaled Medications
Polymer Classification: Architecture
Transformers
The iron core has a substantial relative permeability. Therefore, the magnetic field lines generated due to the current in one winding are almost entirely confined within the core, such that the same magnetic flux permeates each turn of both...
Transformation

