Automatic liver segmentation from abdominal imagesis challenging on the aspects of segmentation accuracy, automation androbustness. There exist many methods of liver segmentation and ways of categorisingthem. Inthis paper, we present a new way of summarizing the latest achievements inautomatic liver segmentation. We categorise a segmentation method according to theimage feature it works on, therefore better summarising the performance of eachcategory and leading to finding an optimal solution for a particularsegmentation task. All themethods of liver segmentation are categorized into three main classes includinggray level based method, structure based method and texture based method. Ineach class, the latest advance is reviewed with summary comments on theadvantages and drawbacks of each discussed approach. Performance comparisonsamong the classes are given along with the remarks on the problems existed andpossible solutions. In conclusion, we point out that liver segmentation isstill an open issue and the tendency is that multiple methods will be employedtogether to achieve better segmentation performance.