1
FOREGROUND EXTRACTION Focussing on: Automatically segment the foreground from an image. Steps following: 1. Low level segmentation (Mean shift, super pixel, watershed) algorithms. 2. After low level segmentation, need to try to give seed points for foreground and background based upon head and torso areas (or) by using face detection method (or) using saliency (or) randomly. 3. After occuring the exact position of the object by using Maximal Similarity Based Region Merging(MSRM) method extraction of the foreground.

Procedure Following

Embed Size (px)

DESCRIPTION

procedure

Citation preview

Page 1: Procedure Following

FOREGROUND EXTRACTION

Focussing on: Automatically segment the foreground from an image.

Steps following:

1. Low level segmentation (Mean shift, super pixel, watershed) algorithms.

2. After low level segmentation, need to try to give seed points for foreground and background

based upon head and torso areas (or) by using face detection method (or) using saliency (or)

randomly.

3. After occuring the exact position of the object by using Maximal Similarity Based Region

Merging(MSRM) method extraction of the foreground.