Handwritten document image binarization: An adaptive K-means based approach

Jana, Pulak Kumar; Ghosh, Soulib; Bera, Suman Kumar; Sarkar, Ram

doi:10.1109/calcon.2017.8280729

Cited by 15 publications

(8 citation statements)

References 16 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Next, these datasheets are scanned using HP flatbed scanner with 300 dpi resolution and the drawn components are extracted automatically from the datasheets. The data are binarized with an appropriate technique as reported in [16] and noise pixels are removed using the method described in [6]. Some sample images are provided the link "https://github.com/Archan462 /Codes".…”

Section: Database Preparationmentioning

confidence: 99%

Offline hand-drawn circuit component recognition using texture and shape-based features

Roy

Bhattacharya

Sarkar

et al. 2020

Multimed Tools Appl

Self Cite

View full text Add to dashboard Cite

Circuit diagram is the very foundation of electrical and electronic sciences. A circuit diagram consists of various symbols called circuit components that specify the functionality of that circuit. Every day-today gadgets that we use are made up with a number of electrical/electronic circuits to play out their particular tasks. Till date circuit designers have to physically enter all data from the hand-drawn circuits into computers, and this procedure requires some investment in terms of time and carries mistakes with high likelihood. To this end, in this paper, we propose a method that relaxes this constraint by introducing a method for recognition of hand-drawn electrical and electronic circuit components, with both analog and digital components included. In the proposed method, the pre-processed images of circuit components are used for training and testing a recognition model using a feature set consisting of a texture based feature descriptor, called histogram of oriented gradients (HOG), and shape based features that include centroid distance, tangent angle, and chain code histogram. In addition, the texture based feature, being large in number compared to others is optimized using a feature selection algorithm called ReliefF. Classification of components is done by using sequential minimal optimization (SMO) classifier. The proposed method has been evaluated on a dataset of 20 different circuit components with 150 samples in each class. The experimental outcome shows that the proposed approach provides average 93.83% accuracy on the present database. We also compare our method with some of the state-of-the-art methods and we see that our method outperforms these methods.

show abstract

Section: Database Preparationmentioning

confidence: 99%

Offline hand-drawn circuit component recognition using texture and shape-based features

Roy

Bhattacharya

Sarkar

et al. 2020

Multimed Tools Appl

Self Cite

View full text Add to dashboard Cite

show abstract

“…The accuracy of the proposed BiGHAM using the feature combination G1B1Bα with an RBF SVM achieved 93.4%, higher than using the existing Bueno's method. Also, if using some existing binarization methods (Sauvola's method [101], K-means based method [102]) instead of the proposed Multi-Region Binarization, the accuracies dropped below 90%, again proved the usefulness of the proposed Multi-Region Binarization for classfication.…”

Section: Discussionmentioning

confidence: 83%

“…A local thresholding binarization method, Multi-Region Binarization, was proposed and both Otsu (a global thresholding binarization method) [24] and Multi-Region binarized images were used in classification as discussed in Sections 5.3, 5.4.1. For comparison, 2 existing local thresholding methods, Sauvola's method [101] and K-means based method [102], were used.…”

Section: Other Local Thresholding Binarization Methodsmentioning

confidence: 99%

“…The binarized images with different binarization methods of some image examples are compared in Figure 5 The testing accuracies using the proposed Multi-Region Binarization were always higher than using the 2 existing binarization methods [101], [102] with either radial basis function (RBF) support vector machine (SVM) or bagging decision trees (BDT) as the classifier, indicating that the shape of the outer patterns on our images also contains information that is useful for classification.…”

Section: Binarized Image Comparisonmentioning

confidence: 98%

“…The deeplearning-based Pedraza's method [73] has a higher accuracy of 93.9% but requires a much higher computational cost (716 MFLOPs), and thus not suitable for portable, real-time applications. Additionally, using 2 existing local thresholding binarization methods [101], [102] instead of the proposed Multi-Region Binarization in the feature combination G1B1Bα, the classification accuracies were less than 90%, proving the usefulness of the proposed Multi-…”

Section: Binarized Image Comparisonmentioning

confidence: 99%

See 2 more Smart Citations

Image processing algorithms for waterborne pathogen classification

Jiawen¹

View full text Add to dashboard Cite

I would like to express my sincere and deep gratitude to my supervisors, Associate Professor SER Wee and Professor LU Yilong, for their invaluable guidance, support, and encouragement. Associate Professor SER Wee provided me with the opportunity to start the research study and has been always approachable and patient to lead me on all aspects of the research like problem defining, data analyzing, method developing, and result presenting throughout my whole PhD study. After his retirement in June 2021, Professor LU Yilong kindly agreed to be my supervisor and gave me much advice and help on the PhD graduation processes including the thesis. The research journey with them has been an unforgettable experience guiding me in my future study and career. I wish to thank the professors and colleagues in the Centre for Bio Devices and Signal Analysis (

show abstract