公开(公告)号：US20230177678A1

公开(公告)日：2023-06-08

申请号：US17806402

申请日：2022-06-10

Applicant: Tata Consultancy Services Limited

Inventor： MANU SHEORAN ,  MEGHAL DANI ,  MONIKA SHARMA ,  LOVEKESH VIG

IPC: G06T7/00 ,  G06V10/25 ,  G06V10/26 ,  G06V10/778 ,  G06V10/44

CPC classification number: G06T7/0012 ,  G06V10/25 ,  G06V10/26 ,  G06V10/778 ,  G06V10/454 ,  G06T2207/10081 ,  G06T2207/30096 ,  G06V2201/07 ,  G06V2201/032

Abstract: State of the art deep network based Universal Lesion Detection (ULD) techniques inherently depend on large number of datasets for training the systems. Moreover, these system are specifically trained for lesion detection in organs of a Region of interest (RoI) of a body. Thus, requires retraining when the RoI varies. Embodiments herein disclose a method and system for domain knowledge augmented multi-head attention based robust universal lesion detection. The method utilizes minimal number of Computer Tomography (CT) scan slices to extract maximum information using organ agnostic HU windows and a convolution augmented attention module for a computationally efficient ULD with enhanced prediction performance.