Encoder-Decoder Variant Analysis for Semantic Segmentation of Gastrointestinal Tract Using UW-Madison Dataset.

in Bioengineering (Basel, Switzerland) by Neha Sharma, Sheifali Gupta, Dalia H Elkamchouchi, Salil Bharany

TLDR

  • The study developed a new method for segmenting the GI tract in MRI images to aid radiation therapy for GI cancer, achieving high accuracy with a Dice value of 0.9082 and IoU value of 0.8796.
  • The proposed method can improve radiation therapy for GI cancer by accurately targeting tumors while avoiding healthy tissues.
  • The study's findings will aid healthcare professionals involved in biomedical image analysis and has potential applications in other medical fields.

Abstract

The gastrointestinal (GI) tract, an integral part of the digestive system, absorbs nutrients from ingested food, starting from the mouth to the anus. GI tract cancer significantly impacts global health, necessitating precise treatment methods. Radiation oncologists use X-ray beams to target tumors while avoiding the stomach and intestines, making the accurate segmentation of these organs crucial. This research explores various combinations of encoders and decoders to segment the small bowel, large bowel, and stomach in MRI images, using the UW-Madison GI tract dataset consisting of 38,496 scans. Encoders tested include ResNet50, EfficientNetB1, MobileNetV2, ResNext50, and Timm_Gernet_S, paired with decoders UNet, FPN, PSPNet, PAN, and DeepLab V3+. The study identifies ResNet50 with DeepLab V3+ as the most effective combination, assessed using the Dice coefficient, Jaccard index, and model loss. The proposed model, a combination of DeepLab V3+ and ResNet 50, obtained a Dice value of 0.9082, an IoU value of 0.8796, and a model loss of 0.117. The findings demonstrate the method's potential to improve radiation therapy for GI cancer, aiding radiation oncologists in accurately targeting tumors while avoiding healthy organs. The results of this study will assist healthcare professionals involved in biomedical image analysis.

Overview

  • The study focuses on segmenting the small bowel, large bowel, and stomach in MRI images using the UW-Madison GI tract dataset to aid radiation therapy for GI cancer.
  • The study tests various combinations of encoders and decoders to identify the most effective method for segmentation.
  • The primary objective is to improve radiation therapy for GI cancer by accurately targeting tumors while avoiding healthy organs.

Comparative Analysis & Findings

  • The study found that the combination of ResNet50 with DeepLab V3+ was the most effective method, achieving a Dice value of 0.9082, IoU value of 0.8796, and model loss of 0.117.
  • The results demonstrate the potential of the proposed method to improve radiation therapy for GI cancer.
  • The study used the UW-Madison GI tract dataset consisting of 38,496 scans to test the various combinations of encoders and decoders.

Implications and Future Directions

  • The study's findings will aid healthcare professionals involved in biomedical image analysis in improving radiation therapy for GI cancer.
  • Future research can explore other combinations of encoders and decoders to further improve segmentation accuracy and potential applications in other medical fields.
  • The proposed method can be tested on other datasets to validate its generalizability and scalability.
Read Full Article