Abstract:Celiac disease is an autoimmune disorder triggered by the consumption of gluten. It causes damage to the villi, the finger-like projections in the small intestine that are responsible for nutrient absorption. Additionally, the crypts, which form the base of the villi, are also affected, impairing the regenerative process. The deterioration in villi length, computed as the villi-to-crypt length ratio, indicates the severity of celiac disease. However, manual measurement of villi-crypt length can be both time-consuming and susceptible to inter-observer variability, leading to inconsistencies in diagnosis. While some methods can perform measurement as a post-hoc process, they are prone to errors in the initial stages. This gap underscores the need for pathologically driven solutions that enhance measurement accuracy and reduce human error in celiac disease assessments. Our proposed method, MeasureNet, is a pathologically driven polyline detection framework incorporating polyline localization and object-driven losses specifically designed for measurement tasks. Furthermore, we leverage segmentation model to provide auxiliary guidance about crypt location when crypt are partially visible. To ensure that model is not overdependent on segmentation mask we enhance model robustness through a mask feature mixup technique. Additionally, we introduce a novel dataset for grading celiac disease, consisting of 750 annotated duodenum biopsy images. MeasureNet achieves an 82.66% classification accuracy for binary classification and 81% accuracy for multi-class grading of celiac disease. Code: <a class="link-external link-https" href="https://github.com/dair-iitd/MeasureNet" rel="external noopener nofollow">this https URL</a>

What problem does this paper attempt to address?

### What problem does this paper attempt to solve? This paper aims to solve the problem of inaccurate measurement of villi and crypts lengths in the diagnosis of celiac disease (CeD). Specifically, the article points out: 1. **Limitations of manual measurement**: Currently, pathologists evaluate the severity of celiac disease by manually measuring the lengths of villi and crypts. However, this method is not only time - consuming but also susceptible to differences among different observers, resulting in inconsistent diagnosis results. 2. **Deficiencies of existing automated methods**: - **Segmentation - based methods**: Although some methods can perform the measurement task as a post - processing step, these methods are prone to errors in the initial stage, especially in generating continuous segmentation masks. - **Lane - detection - based methods**: Traditional lane - detection models are less effective in dealing with villi and crypts with highly variable directions and complex curvatures. - **Line - segment - detection - based methods**: These methods are usually used to detect edges, intersections or region proposals, but are not suitable for detecting objects such as villi and crypts. 3. **Lack of pathology - driven solutions**: Existing automated methods fail to fully incorporate pathological knowledge, such as the feature that crypts are located between the villi shoulders and the crypt boundaries, resulting in low measurement accuracy. To solve the above problems, this paper proposes a new method - **MeasureNet**, which is a deep - learning model based on a polygon - detection framework, aiming to improve the accuracy of villi and crypts length measurement and reduce human error. In addition, the authors also introduce a new dataset **CeDeM**, which contains 750 annotated duodenal biopsy images to support automated celiac - disease - detection research. ### Main contributions of MeasureNet - **Proposing a new task**: Performing celiac - disease detection through polygon detection instead of traditional classification methods. - **Designing the MeasureNet framework**: Combining localization loss and object - driven loss, specifically for accurate measurement. - **Utilizing segmentation masks as auxiliary information**: By mixing strong and weak segmentation - mask features, enhancing the robustness of the model. - **Providing a new dataset CeDeM**: Containing 6,800 annotated villi and crypts, supporting more in - depth research. Through these improvements, MeasureNet significantly outperforms existing methods in multiple indicators, including localization accuracy, measurement error and classification accuracy.

MeasureNet: Measurement Based Celiac Disease Identification