Automatic Detection of Osteosarcoma Based on Integrated Features and Feature Selection Using Binary Arithmetic Optimization Algorithm

Abstract

Osteosarcoma is one of the most common malignant bone tumor mostly found in children and teenagers. Manual detection of osteosarcoma requires expertise and is a labour-intensive process. If detected on time, the mortality rate can be reduced. With the advent of new technologies, automatic detection systems are used to analyse and classify images obtained from different sources. Here, we propose an automatic detection system Integrated Features-Feature Selection Model for Classification (IF-FSM-C) that detect osteosarcoma from the high-resolution whole slide images (WSIs). The novelty of the proposed approach is the use of integrated features obtained by fusion of features extracted using traditional handcrafted feature extraction techniques and deep learning models. It is quite possible that the integrated features may contain some redundant and irrelevant features which may unnecessarily increases the computation time and leads to wastage of resources. To avoid this, we perform feature selection (FS) before giving the integrated features to the classifier. To perform feature selection, we propose two binary variants of recently proposed Arithmetic Optimization Algorithm (AOA) known as BAOA-S and BAOA-V. The selected features are given to a classifier that classifies the WSIs into Viable tumor (VT), Non-viable tumor (NVT) and non-tumor (NT). Experiments are performed and the results prove the superiority of the proposed IF-FSM-C that uses integrated features and feature selection in classifying WSIs as compared to the classifiers which use handcrafted or deep learning features alone as well as state-of-the-art methods for osteosarcoma detection.

Auxiliary Segmentation Method of Osteosarcoma MRI Image Based on Transformer and U-Net

“…It establishes the groundwork for automating the deep learning algorithms' extraction of tumor prediction maps from raw images. Bansal et al [ 56 ] implemented an automatic detection system based on the F-FSM-C classification model. The model can classify the original image into three types: surviving tumor, nonsurviving tumor, and nontumor, reducing the number of network features.…”