FIELD: information technology.
SUBSTANCE: present invention discloses method and device for image segmentation relating to image processing. Image segmentation method includes steps of: image characteristics model is established; foreground test point and background test point of image are produced according to features model; foreground and background classification model are established according to characteristics model, as well as foreground and background test points; image is segmented according to preset graph section algorithm, and specific image is segmented according to preset graph section algorithm by foreground and background classification model and edge information relating to pixels. Foreground and background classification model is established by automatic determination of foreground test point and background test point and by combination of characteristics model, and image segmentation is achieved by foreground and background classification model.
EFFECT: higher efficiency of segmentation of large number of images.
19 cl, 9 dwg
| Title | Year | Author | Number |
|---|---|---|---|
| RELEVANCY ASSESSMENT FOR ARTIFICIAL NEURAL NETWORKS | 2015 |
|
RU2703343C2 |
| BODY MEASUREMENT | 2013 |
|
RU2635226C2 |
| METHOD AND DEVICE FOR ALARM | 2015 |
|
RU2629469C1 |
| METHOD OF CALCULATING MOVEMENT WITH OCCLUSION CORRECTIONS | 2012 |
|
RU2517727C2 |
| METHOD AND APPARATUS FOR IDENTIFYING PORTIONS OF FRAGMENTED MATERIAL WITHIN AN IMAGE | 2016 |
|
RU2694021C1 |
| METHOD FOR RECOGNITION OF GESTURES IN A SERIES OF STEREO FRAMES | 2001 |
|
RU2280894C2 |
| METHOD AND EQUIPMENT FOR IMAGE ENCODING/DECODING | 2019 |
|
RU2752011C1 |
| ATM MANAGEMENT SYSTEM AND METHOD | 2016 |
|
RU2708422C1 |
| WEEDS IDENTIFICATION IN NATURAL ENVIRONMENT | 2017 |
|
RU2735151C2 |
| IMAGE SEGMENTATION METHOD | 2011 |
|
RU2489752C2 |
