Hybrid Deep Learning Model for Endoscopic Lesion Detection and Classification Using Endoscopy Videos

Abstract

In medical imaging, the detection and classification of stomach diseases are challenging due to the resemblance of different symptoms, image contrast, and complex background. Computer-aided diagnosis (CAD) plays a vital role in the medical imaging field, allowing accurate results to be obtained in minimal time. This article proposes a new hybrid method to detect and classify stomach diseases using endoscopy videos. The proposed methodology comprises seven significant steps: data acquisition, preprocessing of data, transfer learning of deep models, feature extraction, feature selection, hybridization, and classification. We selected two different CNN models (VGG19 and Alexnet) to extract features. We applied transfer learning techniques before using them as feature extractors. We used a genetic algorithm (GA) in feature selection, due to its adaptive nature. We fused selected features of both models using a serial-based approach. Finally, the best features were provided to multiple machine learning classifiers for detection and classification. The proposed approach was evaluated on a personally collected dataset of five classes, including gastritis, ulcer, esophagitis, bleeding, and healthy. We observed that the proposed technique performed superbly on Cubic SVM with 99.8% accuracy. For the authenticity of the proposed technique, we considered these statistical measures: classification accuracy, recall, precision, False Negative Rate (FNR), Area Under the Curve (AUC), and time. In addition, we provided a fair state-of-the-art comparison of our proposed technique with existing techniques that proves its worthiness.

Prediction of Early Visual Outcome of Small-Incision Lenticule Extraction (SMILE) Based on Deep Learning

“…After that, LASSO-penalized feature selection was used to identify the most significant features. Then, seven classic machine learning classifiers (Decision-Tree, Extra-Tree, KNN, LightGBM, Random-Forest, SVM, and XGBoost) combined with tenfold cross-validation were applied to train models for predicting each video's classifications [13][14][15]. Two independent prospective SMILE cohorts were processed in the same way to establish the accuracy and robustness of the model in clinical application.…”

Analysis of Colorectal and Gastric Cancer Classification: A Mathematical Insight Utilizing Traditional Machine Learning Classifiers

“…Minimizing entropy aligns with a uniform distribution, promoting randomness. KL-Divergence KL(p||q) gauges the closeness of p to a uniform distribution q, as in Equation (64). ID3 Algorithm: The ID3 algorithm stops tree-building when all labels are the same or no more attributes can split further.…”

“…The research conducted by (Ayyaz et al, 2022) [64] achieved outstanding results in stomach cancer detection, with a remarkable accuracy of 99.80%. They employed various preprocessing techniques, including resizing, contrast enhancement, binarization, and filtering.…”

Deep Feature Fusion and Optimization-Based Approach for Stomach Disease Classification

“…Finally, the selected features were classified using a multi-layer perceptron, and least square SVM. Ayyaz et al [ 25 ] utilized deep CNN models for the recognition of gastric abnormalities. First, they implemented the transfer-learning technique on AlexNet and VGG19.…”