Processing

Please wait...

Settings

Settings

Goto Application

1. WO2022005670 - IMAGE SEGMENTATION TRAINING WITH CONTOUR ACCURACY EVALUATION

Publication Number WO/2022/005670
Publication Date 06.01.2022
International Application No. PCT/US2021/035026
International Filing Date 28.05.2021
IPC
G06K 9/50 2006.1
GPHYSICS
06COMPUTING; CALCULATING OR COUNTING
KRECOGNITION OF DATA; PRESENTATION OF DATA; RECORD CARRIERS; HANDLING RECORD CARRIERS
9Methods or arrangements for reading or recognising printed or written characters or for recognising patterns, e.g. fingerprints
36Image preprocessing, i.e. processing the image information without deciding about the identity of the image
46Extraction of features or characteristics of the image
50by analysing segments intersecting the pattern
G06N 20/20 2019.1
GPHYSICS
06COMPUTING; CALCULATING OR COUNTING
NCOMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS
20Machine learning
20Ensemble learning
CPC
G06K 9/6256
GPHYSICS
06COMPUTING; CALCULATING; COUNTING
KRECOGNITION OF DATA; PRESENTATION OF DATA; RECORD CARRIERS; HANDLING RECORD CARRIERS
9Methods or arrangements for reading or recognising printed or written characters or for recognising patterns, e.g. fingerprints
62Methods or arrangements for recognition using electronic means
6217Design or setup of recognition systems and techniques; Extraction of features in feature space; Clustering techniques; Blind source separation
6256Obtaining sets of training patterns; Bootstrap methods, e.g. bagging, boosting
G06T 2207/10024
GPHYSICS
06COMPUTING; CALCULATING; COUNTING
TIMAGE DATA PROCESSING OR GENERATION, IN GENERAL
2207Indexing scheme for image analysis or image enhancement
10Image acquisition modality
10024Color image
G06T 2207/20081
GPHYSICS
06COMPUTING; CALCULATING; COUNTING
TIMAGE DATA PROCESSING OR GENERATION, IN GENERAL
2207Indexing scheme for image analysis or image enhancement
20Special algorithmic details
20081Training; Learning
G06T 7/12
GPHYSICS
06COMPUTING; CALCULATING; COUNTING
TIMAGE DATA PROCESSING OR GENERATION, IN GENERAL
7Image analysis
10Segmentation; Edge detection
12Edge-based segmentation
G06V 10/44
G06V 10/774
Applicants
  • SONY GROUP CORPORATION [JP]/[JP]
  • SONY PICTURES ENTERTAINMENT INC. [US]/[US]
Inventors
  • CHEN, Mengyu
  • ZHU, Miaoqi
  • TAKASHIMA, Yoshikazu
  • CHAO, Ouyang
  • DE LA ROSA, Daniel
  • LAFUENTE, Michael
  • SHAPIRO, Stephen
Agents
  • LEE, Samuel S.
  • CATANESE, Mark W.
Priority Data
17/179,06118.02.2021US
63/047,75002.07.2020US
Publication Language English (en)
Filing Language English (EN)
Designated States
Title
(EN) IMAGE SEGMENTATION TRAINING WITH CONTOUR ACCURACY EVALUATION
(FR) APPRENTISSAGE DE SEGMENTATION D'IMAGE AVEC ÉVALUATION DE LA PRÉCISION DU CONTOUR
Abstract
(EN) Improving the accuracy of predicted segmentation masks, including: extracting a ground-truth RGB image buffer and a binary contour image buffer from a ground-truth RGB image container for segmentation training; generating predicted segmentation masks from the ground-truth RGB image buffer; generating second binary contours from the predicted segmentation masks using a particular algorithm; computing a segmentation loss between manually-segmented masks of the ground-truth RGB image buffer and the predicted segmentation masks; computing a contour accuracy loss between contours of the binary contour image buffer and the binary contours of the predicted segmentation masks; computing a total loss as a weighted average of the segmentation loss and the contour accuracy loss; and generating improved binary contours by compensating the contours of the binary contour image buffer with the computed total loss, wherein the improved binary contours are used to improve the accuracy of the predicted segmentation masks.
(FR) L'amélioration de la précision de masques de segmentation prédits comprend : l'extraction d'un tampon d'image RVB de réalité de terrain et d'un tampon d'image de contour binaire à partir d'un contenant d'image RVB de réalité de terrain pour l'apprentissage de segmentation ; la génération de masques de segmentation prédits à partir du tampon d'image RVB de réalité de terrain ; la génération de seconds contours binaires à partir des masques de segmentation prédits à l'aide d'un algorithme particulier ; le calcul d'une perte de segmentation entre des masques segmentés manuellement du tampon d'image RVB de réalité de terrain et des masques de segmentation prédits ; le calcul d'une perte de précision du contour entre les contours du tampon d'image de contour binaire et les contours binaires des masques de segmentation prédits ; le calcul d'une perte totale en tant que moyenne pondérée de la perte de segmentation et de la perte de précision du contour ; et la génération de contours binaires améliorés par compensation des contours du tampon d'image de contour binaire avec la perte totale calculée, les contours binaires améliorés étant utilisés pour améliorer la précision des masques de segmentation prédits.
Related patent documents
Latest bibliographic data on file with the International Bureau