In　this　paper,　a　new　robust　improved　Chan-Vese　(ICV)　model　is　proposed　for　image　segmentation,　which　is　built　based　on　the　techniques　of　curve　evolution,　signed　pressure　force　(SPF)　function　and　level　set　method.　Compared　with　the　ICV　model,　the　proposed　method　is　more　robust　to　the　location　of　the　initial　contour.　Similar　to　the　ICV　model,　a　Gaussian　regularizing　level　set　method　(GRLSM)　is　used　to　reduce　the　computational　cost.　Experimental　results　on　some　synthetic　and　real　images　show　that　our　model　is　efficiency.　Moreover,　comparisons　with　the　ICV　model　show　that　our　model　is　more　robust　to　the　location　of　the　initial　contour.　©　2011　IEEE.