Presentation is loading. Please wait.

Presentation is loading. Please wait.

1 Document Image Matching Based on Component Blocks Fuhui Long, Hanchuan Peng, Zheru Chi, and Wanchi Siu Center for Multimedia Signal Processing, Department.

Similar presentations


Presentation on theme: "1 Document Image Matching Based on Component Blocks Fuhui Long, Hanchuan Peng, Zheru Chi, and Wanchi Siu Center for Multimedia Signal Processing, Department."— Presentation transcript:

1 1 Document Image Matching Based on Component Blocks Fuhui Long, Hanchuan Peng, Zheru Chi, and Wanchi Siu Center for Multimedia Signal Processing, Department of Electronic & Information Eng., The Hong Kong Polytechnic Univ., Email: {fhlong, phc, enzheru}@eie.polyu.edu.hk http://pandora.rad.jhu.edu/~phc/publications.html

2 2 Outline Introduction Component Block & Data Structure Matching Algorithm Experiments Discussion & Conclusion

3 3 Document Image Matching Key technique for document image registration & retrieval Can be applied widely for office automation, digital library, video- conferencing, etc.

4 4 Current Techniques Existing methods are mainly based on local features of document page image Cesarini ’ s form-reader system (attributed relational graphs) Shimotsuji ’ s cell struct based 2-dimensional hash table Watanabe ’ s blank form structure of repetitions and positions of cells Tseng and Chen ’ s line segments based method Fan and Chang ’ s line crossing relationship matrix Watanabe and Huang ’ s predefined logical structure for business cards Safari ’ s projective geometry method etc

5 5 Our Approach Decompose a document page image into local component blocks Propose measurements to combine ------local block information ------global page layout information It is closely related to our e-Doc technique, which is developed for document databases

6 6 e-Doc Documents on Papers Image Acquiring Optical Images Pre-Processing Component Block List e-Doc Page Block Organization Functions for Applications The Block-oriented e-Doc technique can be very useful for document databases related applications, including the document page image retrieval, etc.

7 7 Preprocessing Noise removing Region based binarization and foreground extraction Correlation based skew correction Image Blocking: Scan from bottom to top and from left to right Use the simplest region growing method (not pixel-by-pixel, but line-by-line, i.e. if there is a pixel on the out boundary of current block, then grow out one more line.)

8 8 Component Block List …… order=1 bound={(250,429),(45,86)} lang_index = 0.4309 type=English We only make use of the block location & size information for matching.

9 9 Matching Algorithm Procedure CBL-MA; {Input: a CBL for the input document image, a handle to a template image database of K TBLs } {Output: the TBL with the minimum distance D to CBL} {Preprocessing: for k=1,…,K, do begin sort the kth TBL by block size (from small to large); end.} {Note: the preprocessing is not a part of this CBL-MA and needs to be done only once beforehand} begin sort the CBL by block size (from small to large); for k=1,…,K, do begin compute D k, which is the distance between CBL and the kth TBL; end; select the TBL with the minimum D k as output; end. (Notation: TBL – Template Block List, CBL – Component Block List)

10 10 Distance Definitions Size Matching: Location Matching: Total Distance:

11 11 Illustration of the Algorithm BA BTBT …… Sequencing with size Component block list Sequenced block list Template block list 1. Matching with size 2. Matching with location

12 12 Experimental Data A large document template data set of 1350 templates. Define 5 subsets with sizes: 50 templates 100 templates 200 templates 500 templates 1350 templates Use computer to generate all test data (deformation images according to these templates)

13 13 Deformation Types Detection Error Block misdetection rate P m Block misaddition rate P a Block Size Variation Block size variation rate P s Block size variation scale S s Block Location Displacement Block displacement rate P d Block displacement scale S d Block Rotation Block rotation rate P r Block rotation angle D r

14 14 Data Examples Template imageDeformation image

15 15 Results for Detection Errors (1) CBL-MA can perform well (r c > 85%) even when 50% blocks in the block list are lost or wrongly added (see the column of P m = 0.5 and P a = 0.5). Even when 80% blocks are lost or added, this algorithm can still produce matching accuracy nearly 60% (see the column of P m = 0.8 and P a = 0.8). (2) CBL-MA is more insensitive to block misaddition than to block misdetection. This is reasonable because when additional blocks are wrongly put into CBL, the original blocks still play, although weaker, roles. On the contrary, the lost information due to block misdetection is non-recoverable.

16 16 Results for Detection Errors P m: Block misdetection rate P a : Block misaddition rate

17 17 Results for Block Size Variation For block size variation, the influence of parameters P s and S s (here we set the same scale factor for both block width and height) on r c is given in Table 2. When blocks expand or shrink greatly, CBL-MA can keep r c above 90% (the fourth row of Table 2). At the same time, even when all blocks have size variations (P s = 1.0), our algorithm can produce a very high matching accuracy of 95%. Note that the latter corresponds to many office automation applications, where S s is not very large, however, most blocks are subject to some degree of size variation, i.e. P s is close to 1.

18 18 Results for Block Size Variation P s : Block size variation rate S s : Block size variation scale

19 19 Results for Block Displacement For block location displacement, the influence of parameters P d and S d on r c is given in Table 3. Evidently CBL-MA is robust to block location variation (r c always larger than 95%).

20 20 Results for Block Displacement P d : Block displacement rate S d: Block displacement scale

21 21 Results for Block Rotation For block rotation, the influence of parameters P r and D r on r c is given in Table 4. For both cases {D r =15 , P r varies from 0.2 to 1.0} and { P r =0.5, D r varies from 5  to 45  }, CBL-MA produces satisfying classification, even when the test images contain strong deformation, e.g. 50% component blocks have at most 45  rotation, or all component blocks have at most 15  rotation. Notice that block rotation will directly lead to the significant change of block sizes.

22 22 Results for Block Rotation P r : Block rotation rate D r : Block rotation angle

23 23 Results for Template Set Size Here under a general setting of parameters {P a =0.2, P m =0.2, P s =0.2, S s =0.2, P d =0.5, S d =0.5, P r =0.5, D r =15  }, we examine the influence of the template image set size on the matching accuracy. All the five template sets are used. For each template set, we independently generate at least 2000 images for testing. The results are listed in Table 5. It is clear that even when the template set size grows to 500, the matching accuracy is satisfying (>80%). For the 1350- template set (Set-E), r c is still around 70%.

24 24 Results for Template Set Size

25 25 Comparison to Other Algorithms It is noticed that the failure in detecting local features (e.g. line-segments) usually immediately results in bad performance of several other algorithms. However, our experiments demonstrate that even when the block information is partially lost or inaccurate, there is no significant performance reduction of CBL-MA.

26 26 Computational Complexity Denote n as the number of blocks in a CBL, m as the number of blocks in a TBL, K as the number of template images in a document image database. When we use quicksort and binary search algorithms, the typical computational complexity is O(nlogn) for CBL sorting O(Kmlogn) for CBL/TBL size matching O(Km(2T C +1)) for CBL/TBL location matching O(K) for CBL/TBL distance calculation O(logK) for finding the minimum distance Totally O((Km+n)logn)

27 27 Application: Column Table Data Extraction This page matching algorithm is then refined and applied to automatic data extraction of column forms. ---All fields in input image differ from each other a lot and the local-feature based approach can not work well. ---Our algorithm is a powerful tool to find out the correct image template, which is used to annotate the image data fields to accomplish the data extraction successfully.

28 28 Conclusions Based on block list (and tree), our algorithm can effectively make use of the local information of each page block and the global information of page layout. We present a method for effective document image matching. The algorithm gives satisfying performance for various image deformations. The algorithm is robust to image distortion, filled-in text, and noises. We report an successful application of our algorithm in column table data auto-reading.


Download ppt "1 Document Image Matching Based on Component Blocks Fuhui Long, Hanchuan Peng, Zheru Chi, and Wanchi Siu Center for Multimedia Signal Processing, Department."

Similar presentations


Ads by Google