Video Motion Magnification For Accurate Vessel And Neurovascular Bundle Localization In Da Vinci Surgery

Minimally invasive surgery is increasingly performed in the operating room because of its fast recovery and less complication.Unfortunately,there are also some drawbacks.Surgeons can perceive vessels and neurovascular bundles through visual and tactile perception in open surgery.In minimally invasive surgery,surgeons manipulate various surgical tools through the visualization provided by endoscopes without any tactile perception.Hence,it is difficult for surgeons to intuitively perceive small pulsatile motion of vessels and neurovascular bundles from surgical field.Therefore,vessels and neurovascular bundles are easily inadvertent injured,resulting in bleeding or even nerve damage which can affect the physiological function of patients.This work proposes a new surgical video motion magnification method to magnify the pulsatile motion of vessels and neurovascular bundles in surgical videos to help surgeons easily and precisely recognize vessels and bundles by their visual system.The proposed method consists of robust hybrid temporal filtering and deep learned spatial decomposition which based on Eulerian video motion magnification.The proposed hybrid temporal filtering approach can significantly magnify pulsatile motion more consistent with reality and simultaneously keep no-pulsatile regions almost identical to original videos.Additionally,integrating hybrid temporal filtering with learning-based spatial decomposition can reduce noise and ring artifacts in magnified videos.We evaluate our method on surgical videos acquired from robotic prostatectomy,with the experimental results demonstrating the effectiveness of our method.