Enhanced tuberculosis diagnosis: microscopic automatic stitching of sputum samples utilizing the SURF feature detector

Abstract

In conventional TB diagnosis, 100-300 field of view (FOV) microscopic fields are to be observed, which might lead to observer fatigue. For both of these tasks, an automatic stitching framework was studied, which extends to conventional feature-based transformations while incorporating feature matching using affine geometry and RANSAC-based homography refinement, which accounts for the unique low-texture morphology and irregular patterns of Mycobacterium tuberculosis in ZN-stained sputum smears. The system was tested on a set of 10 overlapping image pairs with a fixed overlap of 30%. Among the evaluated image pairs, the proposed optimized method achieved a 100% success rate. Objective zero-pixel metric-based quantitative analysis also validated a higher quality of transparency as compared to other methods. The proposed SURF implementation reached a minimum number of 345.263 zero-pixels, outperforming standard SURF (964.247) and SIFT (1.069.687). This improved robustness to rotation and illumination variations rendered the optimized SURF-affine framework a preferred choice for automatic TB diagnosis systems.