A Novel System for Multi-level Crohn’s Disease Classification and Grading Based on a Multiclass Support Vector Machine

Abstract Background There is lack of real-world data for disease behavior and surgery of Crohn’s disease (CD) from large-scale Chinese cohorts. Methods Hospitalized patients diagnosed with CD in our center were consecutively included from January 2000 to December 2018. Disease behavior progression was defined as the initial classification of B1 to the progression to B2 or B3. Clinical characteristics including demographics, disease classification and activity, medical therapy, development of cancers, and death were collected. Results Overall, 504 patients were included. Two hundred and thirty one (45.8%) patients were initially classified as B1; 30 (13.0%), 71 (30.7%), and 95 (41.1%) of them had disease progression at the 1-year follow-up, 5-year follow-up, and overall, respectively. Patients without location transition before behavior transition were less likely to experience behavior progression. However, patients without previous exposure to a corticosteroid, immunomodulator, or biological agent had a greater chance of experiencing behavior progression. When the long-term prognosis was evaluated, 211 (41.9%) patients underwent at least one CD-related surgery; 108 (21.4%) and 120 (23.8%) of these patients underwent surgery before and after their diagnosis, respectively. An initial classification as B1, no behavior transition, no surgery prior to diagnosis, and previous corticosteroid exposure during follow-up were associated with a lower risk of undergoing surgery. Conclusions This study depicts the clinical features and factors associated with behavior progression and surgery among hospitalized CD patients in a Chinese center. Behavior progression is associated with a higher probability of CD-related surgery, and strengthened therapies are necessary for them in the early phase.

Download Full-text

Analysis and Classification of ECG Signal based on Whale Optimization with Compact Multiclass Support Vector Machine

10.1109/stcr51658.2021.9589026 ◽

2021 ◽

Author(s):

Ganesh Babu C ◽

Gowri Shankar M ◽

Priyanka G S ◽

Bhavani K ◽

Dharun D

Keyword(s):

Support Vector Machine ◽

Support Vector ◽

Ecg Signal ◽

Whale Optimization ◽

Multiclass Support Vector Machine

Download Full-text

MULTI-LEVEL DOCUMENT IMAGE SEGMENTATION USING MULTI-LAYER PERCEPTRON AND SUPPORT VECTOR MACHINE

International Journal of Pattern Recognition and Artificial Intelligence ◽

10.1142/s0218001412530023 ◽

2012 ◽

Vol 26 (06) ◽

pp. 1253002 ◽

Cited By ~ 1

Author(s):

YAN ZHANG ◽

BIN YU ◽

HAI-MING GU

Keyword(s):

Support Vector Machine ◽

Image Segmentation ◽

Research Area ◽

Document Image ◽

Support Vector ◽

Multi Layer Perceptron ◽

Document Image Analysis ◽

Important Research Area ◽

Multi Level ◽

Document Image Segmentation

Document image segmentation is an important research area of document image analysis which classifies the contents of a document image into a set of text and non-text classes. Previous existing methods are often designed to classify text and halftone therefore they perform poorly in classifying graphics, tables and circuit, etc. In this paper, we present a robust multi-level classification method using multi-layer perceptron (MLP) and support vector machine (SVM) to segment the texts from non-texts and thereafter classify them as tables, graphics and halftones. This method outperforms previously existing methods by overcoming various issues associated with the complexity of document images. Experimental results prove the effectiveness of our proposed method. By virtue of our multi-level classification approach, the text components, halftone components, graphic components and table components are accurately classified respectively which would highly improve OCR accuracy to reduce garbage symbols as well as increase compression ratio thereafter simultaneously.

Download Full-text