Conference Paper

Improved Quadtree Method for Split Merge Image Segmentation

M.I.T., Ujjain
DOI: 10.1109/ICETET.2008.145 Conference: Emerging Trends in Engineering and Technology, 2008. ICETET '08. First International Conference on
Source: IEEE Xplore


Image segmentation is one of the important steps in Image processing. This paper introduces an improved quadtree method (IQM) for split-merge called as neighbour naming based image segmentation method (NNBISM) in Kelkar, D. and Grupta, S., (2008), where top-down and bottom-up approaches of region based segmentation techniques are chained. IQM mainly composed of splitting image, onitializing neighbour list and then merging splitted regions. First step uses quadtree for representing splitted Image. In second step neighbour list of every quadtree node, is populated using neighbour naming method (NNM). NNM works at region level, and leads to fast initialisation of adjacency information thus improving the performance of IQM for split merge image segmentation. This populated list is basis for third step which is decomposed in two phases, in-house merge and ginal merge. This decomposing reduces problems involved in handling lengthy neighbour list during merging process .

139 Reads
  • Source
    • "Region splitting and merging techniques [13] [14] [15] starts with splitting an image into small regions and continued till regions with required degree of homogeneity are formed. "
    [Show abstract] [Hide abstract]
    ABSTRACT: In this paper, we present a region growing technique for color image segmentation. Conventional image segmentation techniques using region growing requires initial seeds selection, which increases computational cost & execution time. To overcome this problem, a single seeded region growing technique for image segmentation is proposed, which starts from the center pixel of the image as the initial seed. It grows region according to the grow formula and selects the next seed from connected pixel of the region. We use intensity based similarity index for the grow formula and Otsu's Adaptive thresholding is used to calculate the stopping criteria for the grow formula. We apply the proposed method to the Berkley segmentation database images and discuss results based on Liu's F-factor that shows efficient segmentation.
  • Source
    • "Graph based segmentation merges neighboring regions if the minimum length of links between them is less than the maximum length of links supporting the two regions, and it needs to compute all the links between neighboring pixels. The quad tree method uses complicated neighbor naming algorithms and it traverses the tree based on the initial 4 possible major neighbors to retrieve all neighbors that may constitutes the initial 4 neighbors (Kelkar and Gupta 2008). When implementing combined methods for image segmentation, the conversion from the output of initial process to the input of the following region merging process elicits overhead. "
    [Show abstract] [Hide abstract]
    ABSTRACT: This paper presents an inversed quad tree merging method for hierarchical high-resolution remote sensing image segmentation, in which bottom-up approaches of region based merge techniques are chained. The image segmentation process is mainly composed of three sections: grouping pixels to form image object/region primitives in imagery using inversed quad tree, initializing neighbor list and region feature variables and then hierarchical clustering neighboring regions. This segmentation algorithm has been tested on the QuickBird images and been evaluated and it exhibits good efficiency over initialization of neighbor list for quad tree node/region primitives. This paper also provides a brief proof of the good efficiency of a sorted merge list which can be viewed as an alternative for dither matrix to randomly distribute region merging pairs which is adopted in e-Cognition. KeywordsInversed quad-tree–Image segmentation–Remote sensing–eCognition–ENVI
    Journal of the Indian Society of Remote Sensing 12/2010; 38(4):686-695. DOI:10.1007/s12524-011-0085-3 · 0.76 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: Automated recognition and modeling of 3D objects located in a construction work environment that are difficult to characterize or are constantly changing is critical for autonomous heavy equipment operation. Such automation allows for accurate, efficient, and autonomous operation of heavy equipment in a broad range of construction tasks by providing interactive background information. This paper presents D object recognition and modeling system from range data obtained from flash LADAR, with the goal of rapid and effective representation of the construction workspace. The proposed system consists of four steps: data acquisition, pre-processing, object segmentation on range images, and D model generation. During the object segmentation process, the split-and-merge algorithm, which separates a set of objects in a range image into individual objects, is applied to range images for the segmentation of objects. The whole process is automatic and is performed in nearly real time with an acceptable level of accuracy. The system was validated in outdoor experiments, and the results show that the proposed D object recognition and modeling system achieves a good balance between speed and accuracy, and hence could be used to enhance efficiency and productivity in the autonomous operation of heavy equipment.
Show more