Uncertainty-guided Multitask Regression Network Aided By Optical Flow For Fully Automated Comprehensive Analysis Of Carotid Artery

Fully automated comprehensive analysis of carotid artery(localization of range of interest(ROI),direct quantitative measurement and segmentation of lumen diameter(CALD)and intimamedia thickness(CIMT),and motion estimation of the carotid wall)is a reliable auxiliary diagnosis of cardiovascular diseases,which relieves physicians from laborious workloads.No work has achieved fully automated comprehensive analysis of carotid artery due to five intractable challenges:(1)The heavy reliance on experienced carotid physicians for the selection of ROI limits fully automated studies.(2)The weak structural information of intima-media thickness increases the difficulty of feature encoding.(3)The radial motion of the carotid wall results in the lack of discriminant features of boundaries.(4)Diseased carotid arteries lose many expression features.(5)Optimal weights of multitask regression are hard to tune manually.In this paper,we propose two novel multi-task regression networks,OF-MSRN and OF-UMRN,to solve the above intractable challenges.OF-MSRN is a preliminary attempt of comprehensive analysis of the carotid artery.We evaluate consistency between forward and backward optical flow to improve the accuracy of motion estimation of the diseased carotid.More importantly,we creatively explore an optical flow auxiliary module to take advantage of the co-promotion of segmentation and motion estimation to overcome the restrictions of the radial motion.On this basis,we further advanced and proposed the OF-UMRN for fully automated comprehensive analysis of the carotid artery.We creatively model homoscedastic uncertainty to automated tune the weights of the segmentation task and the direct quantitative measurement task optimally.We take full advantage of the pathological relationships between multiple objects and co-promotion between multiple tasks.We conducted extensive experiments on 202 ultrasound sequences from 101 patients and demonstrated the superior performance of our proposed network for automated comprehensive analysis of carotid arteries.Therefore,the uncertainty-guided multitask regression network aided by optical flow for fully automated comprehensive analysis of carotid artery has a good clinical application prospect in the diagnosis and evaluation of cardiovascular diseases.