For period four we were asked to create 3D models of objects we have used during Covid19 and which carry a special meaning for us, mainly during this period of lockdown and quarantine. We started with some theoretical knowledge concerning the Structure from Motion process and we applied our knowledge on Agisoft Metashape, a platform where we can upload 2D photos of an object and use them to produce the 3D model of it. After taking photos from all angles around my object in a DIY light box, I uploaded them on the platform and started the processing. Even though the steps we had to follow were easy, to achieve a satisfying result at the end appeared to be very difficult. There were many limitations with regard to the material of the object, as for example shiny parts or sharp edges are not easily captured by the algorithm. This created many problems to my object at the stage of the alignment of photos, as some parts of my wooden puzzle were very shiny. At this point, creating masks was essential, given that it helps the system to identify overlapping photos easily and do the alignment successfully. Even though I created masks in all my photos one by one, in the next steps “build dense cloud” and “build mesh”, my object appeared to have many holes I had to close, while the texture of the object seemed as “melted”. In the end, I chose to try with a new pyramid puzzle that is less shiny and I took pictures this time without the light box, just by creating a “hemisphere” of images around my object. Given that the quality of the dataset was very high and that there were not shiny parts, I decided not to create masks. The result appeared to be very satisfying this time, proving that the quality of the photoset is essential.