Abstract
Abstract An image segmentation approach by improved watershed partition and DCT energy compaction has been proposed in this paper. The proposed energy compaction, which expresses the local texture of an image area, is derived by exploiting the discrete cosine transform. The algorithm is a hybrid segmentation technique which is composed of three stages. First, the watershed transform is utilized by preprocessing techniques: edge detection and marker in order to partition the image in to several small disjoint patches, while the region size, mean and variance features are used to calculate region cost for combination. Then in the second merging stage the DCT transform is used for energy compaction which is a criterion for texture comparison and region merging. Finally the image can be segmented into several partitions. The experimental results show that the proposed approach achieved very good segmentation robustness and efficiency, when compared to other state of the art image segmentation algorithms and h...
Highlights
In the literature, segmentation has been known as the process of dividing an image into a subset of connected regions based on an application defined criteria
Among the different segmentation methods, unsupervised image segmentation algorithms have been widely applied for its generalization, which can be broadly divided in three categories: region-based, graph-based and model-based ones
Our work focuses on the region-based algorithm, and a novel approach is proposed, which combines watershed and region merging by intensity and texture extraction
Summary
Segmentation has been known as the process of dividing an image into a subset of connected regions based on an application defined criteria. Many research efforts have been done regarding graph-based image segmentation algorithms [7,8,9] The model based methods strive to capture the underlying structure of the texture and classify image pixels into different segmented regions [10,11,12]. Our work focuses on the region-based algorithm, and a novel approach is proposed, which combines watershed and region merging by intensity and texture extraction. The main drawback of Gabor filtering is that the excessive computational effort can‟t be avoided To address these limitations, the discrete cosine transform is adopted for texture representation, resulting in a fast segmentation and energy compaction approach. Merge I1 by statistical characteristic Image I2 composed of less than 100 regions
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have