Practice at our Center shows that approach using 3D surface imaging for morphometric comparison of patterned injuries does not always lead to accurate conclusions. We decided to evaluate whether a selection protocol focused on analysis phase could enable us to form an early assessment of the outcome of a comparison process, and then to select lesions likely to lead to a probative conclusion. 23 blunt objects were used to create 65 patterned injuries on an experimental model simulating human skin. A blinded analysis and a comparison were conducted on photographs and 3D models of the lesions. Statement of analysis phase was consistent with comparison results in most cases, enabling correct identification of the responsible object or at least keeping it as possibly responsible among 2 to 3 objects. Our protocol has been demonstrated to improve ability to exploit patterned injuries from surface imaging, despite certain limiting factors.