Related Experiment Video
Updated: Jun 24, 2025
04:48
Application of Deep Learning-Based Medical Image Segmentation via Orbital Computed Tomography
Published on: November 30, 2022
2.7K
Enhancing bladder cancer diagnosis through transitional cell carcinoma polyp detection and segmentation: an
Mahdi-Reza Borna1, Mohammad Mehdi Sepehri1, Pejman Shadpour2
1Department of IT Engineering, Faculty of Industrial and Systems Engineering, Tarbiat Modares University, Tehran, Iran.
Frontiers in Artificial Intelligence
|June 14, 2024
Summary
Deep learning models can accurately segment transitional cell carcinoma (TCC) polyps in bladder cancer diagnosis. These models show promise even with low-quality cystoscopy images, aiding early detection without altering clinical workflows.
Area of Science:
- Medical Imaging
- Artificial Intelligence
- Oncology
Background:
- Bladder cancer, particularly transitional cell carcinoma (TCC) polyps, poses a global health challenge.
- Precise segmentation of TCC polyps in cystoscopy images is vital for early diagnosis and treatment.
- Deep learning (DL) offers a promising approach to enhance TCC polyp segmentation.
Purpose of the Study:
- To evaluate the efficacy of various DL architectures for segmenting TCC polyps in cystoscopy images.
- To assess the performance of DL models trained on low-quality datasets.
Main Methods:
- Three DL architectures were evaluated: Unetplusplus_vgg19, Unet_vgg11, and FPN_resnet34.
- Models were trained on a dataset of annotated, low-quality cystoscopy images.
Main Results:
- Unetplusplus_vgg19 and FPN_resnet34 demonstrated promising precision rates of 55.40% and 57.41%, respectively.
- These performance levels suggest suitability for clinical integration without workflow disruption.
Conclusions:
- DL models show significant potential for TCC polyp segmentation, even with lower-quality imaging data.
- The findings indicate that DL can improve the timeliness of bladder cancer diagnosis while maintaining current clinical processes.
