The development of quantitative digital phenotyping methods for evaluation of wild potato (section Petota Dumort., genus Solanum L.) tuberization is required for annotation of genebank collections and selection of the suitable donor material for potato breeding. There are no available methods specifically designed for the quantitative analysis of wild potato tuber morphology. The current study is devoted to evaluation of wild potato tubers’ morphological characteristics using a digital image processing technique. For this purpose, the mobile application SeedSounter developed previously for grain analysis was specifically adapted for tuber phenotyping. The application estimates the number and shape of objects scattered on a standard sheet of white paper (i. e. A3 or A4). Twelve accessions from the VIR genebank collection belonging to nine Petota species were grown in pots protected with garden fabric during the growing season of cultivated potato (Novosibirsk region). Tubers were collected form plants of nine genotypes. Three genotypes did not produce tubers. The weight of tubers collected from each plant was measured. The tuber yield from each plant was analyzed using SeedCounter (http://wheatdb.org/seedcounter). The number of tubers per plant was counted; the following characteristics were extracted from the images of individual tubers: length, width, projected area, length to width ratio, сircularity, roundness, rugosity and solidity. One-way ANOVA showed a significant effect of genotype on all measured characteristics. A pairwise comparison of nine Petota accessions using all measured parameters revealed statistically significant differences between 86 % of pairs. The overall tuber yield volume for each plant was calculated as a sum of volumes of individual tubers; tuber volume was calculated from its length to width ratio and projected area. A strong correlation between the evaluated tuber yield volume and yield weight was shown. We propose tuber yield volume as a characteristic for a general evaluation of tuberization for wild potato, implementing the four-step scale from 0 to 3. According to this characteristic, the twelve wild potato accessions studied could be divided into four groups with different tuberization abilities. The evaluated tuberization ability is partially in accordance with previously obtained VIR data. The results presented demonstrate the possibility to use SeedCounter for wild potato collections phenotyping.