Image analysis has been used in medicine and veterinary science as an increasingly important auxiliary tool for histopathology evaluation, because of its potential for decreasing the intrinsic subjectivity of the human visual analysis. However, the geometrical features currently used are often inefficient, not always leading to correct diagnosis. The present work has as its main objective to suggest several approaches to geometric and texture characterization of the cell nuclei, characterized by low computational cost, rapid analysis and sound biological interpretation and good agreement with the visual analysis by expert pathologists. We consider digital images of histological slides obtained from histiocytoma, mastocytoma and transmissible venereal tumors in dogs. Segmented cell nuclei of these tumors were evaluated morphometrically considering the features normally used in medicine together with novel features suggested in the present work, applied in order to characterize the distribution and morphology of the nucleoli and heterochromatin inside the nuclei. Statistical characterization of the obtained results was performed through variance and discriminant analysis, leading to the identification of the most important features. The approach proposed in this article has been verified to be an effective auxiliary tool for differential diagnosis of histiocytoma, mastocytoma and TVT in dogs. Its low computational cost allows fast morphometric evaluation, paving the way for real-time tumor diagnosis.