Abstract: Disclosed herein are systems and methods for selecting re-usable car wheels, comprising, receiving one or more images of an evaluated used car wheel, transforming one or more of the images to multi-dimension feature vectors, mapping the evaluated used car wheel to one of a plurality of used car wheels by applying one or more trained ML models to classify the multi-dimension feature vector. The one or more trained ML models are trained using a plurality of multi-dimension feature vectors each created based on a plurality of images of a plurality of used car wheels, computing a price value of the mapped used car wheel based on a plurality of sale transactions of used car wheels such as the mapped used car wheel, and generating instructions to indicate whether the evaluated used car wheel is re-usable or not based on comparison between the computed price value and a price threshold.