Precise segmentation of objects is an important problem in tasks like class-agnostic object proposal generation or instance segmentation . The refinement utilizes superpixel pooling for feature extraction and a novelsuperpixel classifier to determine if a high precision superpixel belongs to anobject or not . Our experiments show an improvement of up to 26.0% in terms of average recall compared to original AttentionMask. Furthermore, qualitative andquantitative analyses of the segmentations reveal significant improvements interms of boundary adherence for the proposed refinement compared to various deep learning-based state-of-the-art object proposal systems. The refinement uses superpixel poolsing for . feature extraction .

Author(s) : Christian Wilms, Simone Frintrop

Links : PDF - Abstract

Code :
Coursera

Keywords : superpixel - refinement - object - proposal - extraction -

Leave a Reply

Your email address will not be published. Required fields are marked *