An intelligent fragile watermarking scheme based on contourlets for effective detection, localization and recovery of tampered regions in handwritten document images
| Autor: | S. Nirmala, K. R. Chetan |
|---|---|
| Rok vydání: | 2017 |
| Předmět: |
business.industry
Computer science Quantization (signal processing) ComputingMethodologies_IMAGEPROCESSINGANDCOMPUTERVISION 020206 networking & telecommunications Pattern recognition Watermark 02 engineering and technology Contourlet 0202 electrical engineering electronic engineering information engineering Curvelet Embedding 020201 artificial intelligence & image processing Artificial intelligence business Quantization (image processing) Digital watermarking |
| Zdroj: | ICACCI |
| DOI: | 10.1109/icacci.2017.8125874 |
| Popis: | Handwritten document images are a special class of document images having a lot of variations in font-style, font size, thickness and interspacing of words and lines. Designing an effective fragile watermarking scheme to protect such images needs capturing both structure and information content which is an open research challenge. A novel contourlet based fragile watermarking technique for optimal detection of tampered locations, localization and recovery of handwritten document images is put forth in this paper. The original image is t subjected to Contourlets transform. The first and second level contourlet coefficients are used as watermark. The watermark is embedded intelligently into robust locations based on the significance of the contribution of the contourlet coefficients. A quantization based embedding technique is used for embedding. The tamper detection is performed by comparing the blocks of generated and extracted watermarks and component labelling technique is used to find out different tampered objects. The recovery of tampered objects is carried out using the watermarks embedded at robust locations. The proposed watermarking scheme using contourlets is compared with a popular technique using curvelets. The proposed technique outperforms in detecting tampered locations and recovery of the tampered information. This work also exhibits better efficiency in terms of PSNR values. Further the proposed method computationally is less expensive compared to the existing method based on curvelets. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|