The purpose of this article was to describe the use of a machine learning model to categorize representative fluoroscopic bladder images of bladders into clusters based on similarity, as well as the 3 shape and 2 texture standard bladder classifications established by urologists which were used as signs that can be indicative of certain urological diseases. A cross sectional study of 417 pediatrics and adolescents was conducted where they had undergone video urodynamics(VUDS), with fluoroscopic images taken in anterior and posterior view, as well as a voiding cystourethrogram(VCUG). Then, the fluoroscopic images were classified independently through two methods: by an unsupervised computer using machine learning through cluster analysis and by clinicians. This study established standardized classifications for bladder shape, such as christmas tree, round, and oblong, and contour, such as smooth and trabeculated. The study also established that vesicoureteral reflux was more associated with bladders with the christmas tree shape and the trabeculated contour compared to the round, oblong, and smooth bladders. However, in contrast with the previous finding, it was also found out that bladder shape did not fully tell the whole story about the degree of bladder dysfunction and there were also other factors at play. Additionally, while the AI did not classify the bladders based on the same terminology as the clinicians, it still grouped bladders into clusters based on their similarity in shape and texture. The study underscores the need for standardized classifications for bladder shape and size to improve the diagnostic value fluoroscopic exposure imagery. The study also explains that although AI is an incredibly useful tool in bladder fluoroscopic imagery analysis, the “black box effect”(the difficulty of interpreting the reasons behind the decisions of deep learning models) holds it back.