An Information Theoretic Framework for Image Segmentation (TA-L1)

Author(s) : Jaume Rigau (Institut d'Informātica i Aplicacions, Spain) Miquel Feixas (Institut d'Informātica i Aplicacions, Spain) Mateu Sbert (Institut d'Informātica i Aplicacions, Spain)

Abstract : In this paper, an information theoretic framework for image segmentation is presented. This approach is based on the information channel that goes from the image intensity histogram to the regions of the partitioned image. It allows us to define a new family of segmentation methods which maximize the mutual information of the channel. Firstly, a greedy top-down algorithm which partitions an image into homogeneous regions is introduced. Secondly, we define an histogram quantization algorithm which clusters color bins in a greedy bottom-up way. Finally, the resulting regions in the partitioning algorithm can optionally be merged using the quantized histogram.

Menu