Skip to Main Content
Interactive machine learning for postprocessing CT images of hardwood logsAuthor(s): Erol Sarigul; A. Lynn Abbott; Daniel L. Schmoldt
Source: Proceedings, 5th International Conference on Image Processing and Scanning of Wood. 177-186.
Publication Series: Miscellaneous Publication
PDF: View PDF (577.83 KB)
DescriptionThis paper concerns the nondestructive evaluation of hardwood logs through the analysis of computed tomography (CT) images. Several studies have shown that the commercial value of resulting boards can be increased substantially if log sawing strategies are chosen using prior knowledge of internal log defects. Although CT imaging offers a potential means of obtaining this knowledge, the automated analysis of the resulting images is difficult, particularly for hardwood species, because of the natural texture/density variations within the wood material. In spite of the difficulties, a few researchers have developed image-analysis systems that demonstrate good statistical accuracy in locating and identifying defects to create "classified" images. Even with good quantitative results, however, classified images can often be improved qualitatively through postprocessing steps that refine the shapes of the detected image regions. To be most effective, postprocessing operations should utilize domain knowledge that is specific to the type and position of different defects. This paper describes an interactive approach for acquiring this domain knowledge. A system has been developed that generates postprocessing rules by observing the operations that are performed as a human user interactively edits a classified CT image. Based on these observations, the system infers rules that can be used subsequently for automatic postprocessing of CT images. The system is incremental, in that the system is capable of updating its rules at a later time. Laboratory tests have shown good improvements to classified images from two red oak logs.
- You may send email to firstname.lastname@example.org to request a hard copy of this publication.
- (Please specify exactly which publication you are requesting and your mailing address.)
- We recommend that you also print this page and attach it to the printout of the article, to retain the full citation information.
- This article was written and prepared by U.S. Government employees on official time, and is therefore in the public domain.
CitationSarigul, Erol; Abbott, A. Lynn; Schmoldt, Daniel L. 2003. Interactive machine learning for postprocessing CT images of hardwood logs. Proceedings, 5th International Conference on Image Processing and Scanning of Wood. 177-186.
- An interactive machine-learning approach for defect detection in computed tomogaraphy (CT) images of hardwood logs
- Robust Spatial Autoregressive Modeling for Hardwood Log Inspection
- Rule-driven defect detection in CT images of hardwood logs
XML: View XML