REVERSIBLE FRAGILE WATERMARKING BASED ON MANHATTAN DISTANCE COORDINATE POINT ON DIGITAL MAP

• Main Author: Nidya Neyman, Shelvie
• Format: Dissertations
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/52317
• Institution: Institut Teknologi Bandung

Digital map has a high data precision, automated process and lossless scaling compared to printed map. The ease of storage and distribution of digital map bring the ease to change as the consequence. This can cause difficulties in the examination of the authenticity of the contents of the map (the integrity assurance of map data). One solution to guarantee the integrity of the data is the digital watermarking, i.e. the insertion of integrity marker into digital media. Further research is needed to find redundancies region on the map for a marker insertion that produces distortion as small as possible. This research aims to produce new watermark insertion technique on reversible fragile watermarking scheme on digital maps. Manhattan distance function is the producer of redundancies region on a digital map where the integrity watermark will be inserted. This function produces a greater distance value or longer trajectory than other distance function, e.g. Euclidean distance, so that the distribution of distortion is more equitable and has smaller influence. The scheme of Reversible Fragile Watermarking in this research was developed to the map vector data model with a line and area spatial feature that work in the realm of spatial domain. The scope of the scheme are: (1) digital watermarking algorithm for the insertion and extraction of the marker on the map, (2) algorithm of cryptographic hash function to generate the integrity marker (watermark), and (3) the parameter establishment of performance measurement of the scheme. Digital watermarking algorithm was developed to meet the watermark characteristics which were susceptible to the manipulation of map media (fragile) and to the recovery capability of the map to its original values (reversible). The generation of this integrity marker used an existed cryptographic hash function, MD5. Other contribution of this research was the proposal of a set of performance measurement parameters of Reversible Fragile Watermarking scheme that can be used for various types of digital media. Those parameters are: the insertion efficacy, invisibility, fidelity, data payload, fragility, blindness, and reversibility. iv The parameter of the insertion efficacy is measured by the percentage of watermark that can be re-extracted; invisibility is the distortion measurement of watermarked-media; fidelity is the correlation measurement between reversible media and the initial media; data payload is the capacity measurement of watermark insertion; fragility is the susceptibility measurement to media manipulation; blindness does not require media/initial watermark; and reversibility is proposed to measure the ability of the recovery message through the difference calculation between reversible media and initial media. Performance level of the scheme in this research can be adjusted through threshold ? variable and digit q determinant. Moreover, the trade-offs between performance-parameters are also needed to be considered in determining the performance that are adjusted to the needs. Increased level of data payload will lower the level of invisibility and reversibility, and vice versa. The trial scheme used 27 pieces of digital map type shapefile of BIG which has varies number of features in scale 1:5.000, 1:10.000, 1:25.000 and 1:250.000 with different threshold value for each scale. The insertion efficacy in the algorithm was set at the level of 100% with the parameter value q equals to six, in accordance with the default shapefile coordinate decimal value. Invisibility parameter indicated that the watermark map distortion was close to zero, the average RMSE values were 0.000000444851 for coordinate X, 0.000000187563 for coordinate Y, and maximum RMSE values were 0.000022997147 coordinate X, 0.000015603411 coordinate Y. This research proposed the addition of a new measurement for the distortion of watermarked map, the calculation of the shiftedcoordinates is in meters. The shift of 5K map was 0.015157 m, 0.049879 m of 10K map, 0.131316 m of 25K map, 0.230608 m of 50K map, 1.783647 m of 250K map. Fidelity level of reversible map reached the average NC correlation value of 0.9999999 or close to 1 between reversible map and original map. Data payload parameter indicated that the number of coordinates on watermarked map ranged between 3.7%-55.59%. The test results of fragility parameter in this research were represented by variable ? as number of features which experienced some attacks. Attacks detected by this scheme were modified value, addition, deletion and modification of feature sequence. Blindness parameter was met by the scheme because the process of watermark extracting and verification only required watermarked map without involving the initial map or initial watermark. Reversibility level showed the 0.005-0.5 meters reversible map. If the threshold value is 10-4 meters for all scale maps, the distortion is in the range of 10-6 -10-7 meters. Reversibility level showed that the map recovery capability was fairly high, in the range of 96.34%- 99.84% of its original state.