An Algorithm to Detect Bone Tumor from Human’s Finger X-Ray Images Using Hybrid Edge Detection Approach

Bone tumors in fingers, though rare, can significantly impair hand function if not detected early. This paper introduces a novel algorithm for detecting bone tumors from human finger X-ray images using a hybrid edge detection approach. The proposed method combines Canny, Sobel, and Laplacian operators to capture both fine and coarse edges, improving the detection of tumor boundaries in complex bone structures. Preprocessing steps include histogram equalization and Gaussian blurring for noise reduction and contrast enhancement. The hybrid edge detection framework enhances structural details by leveraging gradient and region- based information, overcoming limitations of single-method detectors. A dataset comprising X- ray images of both healthy and tumor-affected fingers was used to validate the algorithm. The proposed system achieved an average accuracy of 91.05%, outperforming traditional methods. Notably, it effectively identified features like image enhancement and edges, which are key tumor identification indicators. The algorithm is computationally efficient, suitable for real-time use in resource-limited clinical settings. Future enhancements include integration with deep learning for classification and expansion to other bone regions. This work offers a promising algorithm for early and accurate diagnosis of bone tumors in human’s finger X-rays images.