Robust Recognition of Planar Shapes Under Affine Transforms Using Principal Component Analysis

Tzimiropoulos, Georgios; Mitianoudis, Nikolaos; Stathaki, Tania

doi:10.1109/lsp.2007.896434

Cited by 14 publications

(3 citation statements)

References 14 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The coordinates of the windows are updated as shown in where L is the original length of the pre-sawing lines and is equation (10). [16] equal to the width or height of the inspected image, and w is the width of the pre-sawing lines.…”

Section: B Steepest Descent Methodsmentioning

confidence: 99%

“…Some This inspection and alignment process in repeated until the approaches have attempted to couple other transformations error satisfies the tolerance level. After the misalignment is with Affine, such as Wavelet [7], MSA [8], cosine [9], and removed, the wafer is transferred to the tool position and it PCA [10]. These approaches are targeted toward recognition is cut by a diamond blade.…”

mentioning

confidence: 99%

See 1 more Smart Citation

A Precise Inspection Technique for Wafer Pre-sawing Lines using Affine Transformation

Kim

Lee

Kim³

et al. 2008

2008 15th International Conference on Mechatronics and Machine Vision in Practice

View full text Add to dashboard Cite

This study proposes a simple and fast technique wafer can be employed as alignment marks. However, they for detecting wafer cutting lines using the Affine transformation are not available sometimes because the size of the marks before sawing. This technique was developed for the purpose is insufficient or their quality is inadequate. In some cases of wafer alignment when alignment marks are unavailable. i their qusablity sinadequate. Is en case The edges of the pre-sawing lines are complex and there are there are no patterns usable patterns on the wafer. When this bright shapes in the sawing lines, so it is difficult to apply happens, we can consider using the intersection of the cutting the conventional edge finding method. The sawing lines are lines. The cutting lines are usually called pre-sawing lines. represented by rectangular windows which are deformed by an The information that should be detected is the misalignment Affine transformation. The amount of deformation is determined in the xyO directions. The pre-sawing lines may be detected when the grey level in the window is a global minimum in the image. The global minimum is found by the golden section search. byledged or blo searchng However, these methods are The algorithm was tested using 13 sample images and repeated influenced by the patterns around the lines and a substantial 10 times for each image. The results showed that the processing processing time is needed to determine the rotational angle, time was below 1200msec and the deviation was in the sub-pixel 0. The result also needs to be highly precise and reliable, level. We found that the proposed method can be applied to align but these methods are not always satisfactory in this respect. wafers in the dicing process. Therefore, this study proposes a method of detecting the pre-

show abstract

Section: B Steepest Descent Methodsmentioning

confidence: 99%

mentioning

confidence: 99%

A Precise Inspection Technique for Wafer Pre-sawing Lines using Affine Transformation

Kim

Lee

Kim³

et al. 2008

2008 15th International Conference on Mechatronics and Machine Vision in Practice

View full text Add to dashboard Cite

show abstract

“…To decrease the sensitivity to noise and local shape deformation, matching is usually done in a transformed domain. Examples of these approaches include the Fourier descriptor (FD) [13], the wavelet descriptor (WD) [2, 14–17] and principal components analysis descriptors [18].…”

Section: Introductionmentioning

confidence: 99%