Image segmentation is an essential phase of medical image processing. Orthopaedics practitioners suggest X-Ray imaging to detect bone related diseases of the patient. Due to increase in the number of bone cancers, bone tumor detection and its segmentation from X-Ray image has become thirst area of research in the medical image analysis. In this research paper, an intelligent assistive algorithm has been proposed, which is called BTDBB to identify bone tumor from human being’s X-Ray images. The proposed algorithm accepts human arm X-Ray images, converts into grayscale, do Gaussian filtering to remove noise and enhancement, Segmentation to divide image into different parts based on threshold, detection of ROI using binary blob pattern analysis, crop image to retain ROI only, measurement of tumor size and finally, detection of bone tumor. The algorithm is implemented in Scilab 5.5.2 open source image processing software using 109 X-Ray images as dataset, out of which 82 images were bone tumor infected. We have compared our proposed algorithm with existing algorithms/methods for performance evaluation using different types of accuracies as evaluation metrics. Further, we have found that proposed algorithm yields an average accuracy of 99.12% and hence, it is superior in performance over existing selected algorithms/methods based on selected parameters.

图像分割是医学图像处理的重要阶段。骨科医生建议使用X射线成像来检测患者的骨相关疾病。由于骨癌数目的增加，从X射线图像进行骨肿瘤检测及其分割已成为医学图像分析领域研究的热点。在这篇研究论文中，提出了一种智能辅助算法，称为BTDBB，它可以从人的X射线图像中识别出骨肿瘤。提出的算法接受人体X射线图像，转换为灰度图像，进行高斯滤波以去除噪声和增强效果，基于阈值进行分割以将图像分为不同部分，使用二进制斑点模式分析检测ROI，仅保留ROI的作物图像，测量肿瘤大小，最后检测骨肿瘤。该算法在Scilab 5.5.2开源图像处理软件中实现，使用109张X射线图像作为数据集，其中82张图像受到了骨肿瘤的感染。我们已将我们提出的算法与使用不同类型的精度作为评估指标进行性能评估的现有算法/方法进行了比较。此外，我们发现，提出的算法产生的平均精度为99.12％，因此，其性能优于基于选定参数的现有选定算法/方法。