Open Access
Recursive Method to Detect and Segment Multiple Rectangular Objects in Scanned Images
Cormac Herley
- 01 Jan 2004
pp 27
TL;DR: In this paper, one dimensional projections of the image were used to recursively simplify the segmentation problem efficiently and robustly, and the algorithm was shown to perform well on real data from user scanned images.
read more
Abstract: Rectangular objects frequently appear in certain classes of images. Scanned images of photographs, receipts or other small objects are obvious examples. It is often desirable to detect the existence, number and location of such rectangular objects to single them out for separate treatment. While image segmentation is generally a hard problem, we find that by considering one dimensional projections of the image it is possible to recursively simplify the segmentation problem efficiently and robustly. We outline the algorithm and its components and discuss its performance on real data from user scanned images.
read more
Chat with Paper
AI Agents for this Paper
Find similar papers on Google Scholar, PubMed and Arxiv
Write a critical review of this paper
Analyze citations of this paper to find unaddressed research gaps
Citations
Efficient inscribing of noisy rectangular objects in scanned images
Cormac Herley
- 24 Oct 2004
TL;DR: This work examines the question of identifying and segmenting noisy rectangles, where edges may be ragged, corners may be missing and so on, and tests the robustness of the scheme on receipts and small documents obtained from real scans.
26
Patent
Method of editing static digital combined images comprising images of multiple objects
Ilya V. Kurilin,Ho-Keun Lee,Ilya Vladimirovich Safonov +2 more
- 18 Jul 2012
TL;DR: In this article, a method of editing a combined image which includes a plurality of object images is presented, which includes displaying an outline rectangle which forms outlines of one or more object images within the combined image, and then renewing displaying of the outline rectangle according the segmenting or combining result.
15
Segmentation of Rectangular Objects Lying on an Unknown Background in a Small Preview Scan Image
Michael Guerzhoy,Hui Zhou +1 more
- 28 May 2008
TL;DR: A method to segment rectangular objects that lie on a slightly textured background of an a-priori unknown colour to detect overlapping and touching objects on a very small-sized preview scan image on a coloured/textured background.
Image processing for drawing recognition
TL;DR: This work proposes application of conducting homography for the video stream which can be obtained from the webcam, and proposes algorithm which could be successfully used for this kind of application.
References
A method of detecting the orientation of aligned components
TL;DR: This report describes a method of detecting the orientation of a set components which are located along parallel lines by using the histogram of nearest neighbor directions.
159
Show-through cancellation in scans of duplex printed documents
TL;DR: An adaptive linear filtering scheme is developed for the electronic removal of show-through using scans of both sides of the document using first physical principles to obtain a simplified mathematical model.
135
Robust detection of skew in document images
A. Chaudhuri,Subhasis Chaudhuri +1 more
TL;DR: A robust yet fast algorithm for skew detection in binary document images based on interline cross-correlation in the scanned image, which does not require prior segmentation of the document into text and graphics regions.
95
Recursive method to extract rectangular objects from scans
Cormac Herley
- 24 Nov 2003
TL;DR: A simple, efficient and robust way to automatically detect that the scanned image consists of many small objects rather than a single large one, and segment appropriately is presented.