<![CDATA[This study develops a web-based steganography model to insert the identity of artisans in the form of palmprint images into the image of gringsing ikat woven cloth as a medium for ownership authentication. The method used in the insertion process combines a Convolutional Neural Network and the Least Significant Bit. In contrast, extracting or re-introducing palmprint images from stego images is carried out using a CNN-based classification model. This system was tested with two scenarios; in the first scenario, one palmprint image was inserted into 26 different cloth motifs, while in the second scenario, one cloth motif was inserted into 99 different palmprint images. The test results showed that the system produced consistent confidence values for all cloth motifs in the first scenario. In contrast, in the second scenario, the system achieved an average confidence of 93.5% and a recognition accuracy of 87%. The developed application has proven to be efficient with a reduction in stego image size of up to 66% while maintaining the quality of the stego image, as well as a speedy average execution time of 0.15 seconds for insertion and 0.09 seconds for extraction. These findings prove that the developed steganography model can effectively insert and re-recognize identity images (palmprints) in woven cloth images and has the potential to be applied as an image-based craft product ownership verification system.]]>
