3D METHOD FOR DETECTING BREAST CANCER USING 3D MAMMOGRAPHY AND APPARATUS USING THE SAME

According to the present invention, a method for detecting breast cancer using 3D mammography comprises the steps of: in the presence of a temporal excitation and aggregation (TEA) block including a motion excitation (ME) operation and a multiple temporal aggregation (MTA) operation, acquiring, by a...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: KIM BO HYUNG, PARK GANG IN, CHO YOON MIN, CHA BYUNG HEE
Format: Patent
Sprache:eng ; kor
Schlagworte:
Online-Zugang:Volltext bestellen
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
Beschreibung
Zusammenfassung:According to the present invention, a method for detecting breast cancer using 3D mammography comprises the steps of: in the presence of a temporal excitation and aggregation (TEA) block including a motion excitation (ME) operation and a multiple temporal aggregation (MTA) operation, acquiring, by a computing device, a video image including a search target; dividing, by the computing device, the video image into T frames; passing, by the computing device, a plurality of TEA blocks using the T frames as input values and acquiring a predetermined feature value; and performing, by the computing device, a region proposal network (RPN) operation using the predetermined feature value as an input value to generate a bounding box at the coordinates determined to be breast cancer among the T frames. Therefore, the present invention detects breast cancer by detecting areas with movement patterns through comparison with previous frames. 본 발명에 따르면, 3D 마모그래피를 이용하여 유방암을 탐지하는 방법에 있어서, ME 연산(Motion Excitation) 및 MTA 연산(Multiple Temporal Aggregation)을 포함하는 TEA(Temporal Excitation and Aggregation) Block이 존재하는 상태에서, 컴퓨팅 장치가, 탐색 대상이 포함된 비디오 영상을 획득하는 단계; 상기 컴퓨팅 장치가, 상기 비디오 영상을 T개의 프레임으로 분할하는 단계; 상기 컴퓨팅 장치가, 상기 T개의 프레임을 입력 값으로 하여 복수의 TEA Block을 통과시키고 소정의 feature 값을 획득하는 단계; 및 상기 컴퓨팅 장치는, 상기 소정의 feature 값을 입력 값으로 RPN 연산(Region Proposal Network)을 수행하여 상기 T개의 프레임 중 상기 유방암이라고 판단된 좌표에 bounding box를 생성하는 단계를 포함하는 것을 특징으로 하는 방법을 제시한다.