Tamper detection, localization and self-recovery using slant transform with DNA encoding for medical images

• The slant transform has the best coefficient compaction among the non-sinusoidal transforms. The 2 × 2 blocks of the DC value of the image after subjecting the slant transform are used for watermark generation, which is a self-concealing type. • Unlike other watermarking schemes, the chaotic sequence is not only used for block mapping; in addition, from a single chaotic sequence, three different indexes are generated, such as the random DNA rule base index, block mapping index, and decimal index. These indexes are used in various steps of watermark generation and watermark encryption to make it difficult for the hacker to decrypt the code of the watermark. • The multi-region attack analysis presents the effect of localized and non-localized tampering ( Section 4.3.4 ) on a medical image with relevant metrics. • To better visualize the work, mathematical modelling of the proposed algorithm’s processes, such as watermark generation, watermark embedding, tamper creation, tamper detection, and tamper localization, is performed. A subset of 8 × 8 images is taken as a reference. This work proposes an improved self-embedding semi-fragile watermarking scheme for a smart healthcare system using telemedicine. The watermark is produced by encoding the Slant Transform (ST) coefficients of each block of the host image. DNA logistics generates the authentication bits for tamper detection. An eight-bit watermark is generated from the coefficients of the host image, which is further encrypted by DNA rules to improve the security of the watermarking scheme. The proposed watermarking scheme can resist content removal, collage, multi-region, and text addition attacks. The host image is self-recovered successfully from the tamper region when the host image is altered up to 50 %. The simulation results emphasize that the precision of tamper detection and localization is 94.95 % and 100 % after level-2 inspection. The proposed watermarking scheme is evaluated based on the average peak signal-to-noise ratio. The results emphasize that the proposed algorithm has better tamper detection, localization, and self-recovery than the existing works without compromising the image's quality.