This study compares the applicability of the alpha shape and convex hull algorithms for determining the geometry of tire footprints obtained from image analysis. The novelty of the study lies in the direct comparison of these two contour-extraction methods using a common image-processing procedure and an engineering-oriented multicriteria evaluation of tire footprint geometry. A total of 20 tire footprint samples acquired under tire pressures ranging from 4 to 7 bar and vertical loads from 100 to 1000 N were analyzed. Image segmentation was performed using automatic Otsu thresholding, followed by contour generation using the alpha shape and convex hull algorithms. The generated contours were evaluated in terms of characteristic dimensions, surface area, number of contour points, contour-point density, overlap and boundary agreement with the segmented footprint used as a proxy reference, and multicriteria assessment. The mean footprint width was 18.72 mm for the alpha shape and 18.93 mm for the convex hull, whereas the corresponding mean heights were 43.08 and 42.26 mm, respectively. Substantial differences were observed in the number of contour points and contour-point density: the alpha shape algorithm produced an average of 257.6 contour points, whereas the convex hull algorithm required only 33.5 points. The proxy-reference analysis showed slightly better local contour agreement for the alpha shape, whereas the convex hull provided a substantially more compact contour representation. After pooled min–max normalization, the multicriteria evaluation yielded a higher score for the convex hull algorithm in 19 of the 20 analyzed cases. The results support the suitability of the convex hull method when the primary objective is compact representation of global footprint dimensions, whereas the alpha shape is preferable when preservation of local boundary features is important. The main limitations are the relatively small dataset, the use of a single alpha parameter value, and the use of the segmented footprint as a proxy rather than a manually validated ground-truth contour.