<![CDATA[The authenticity and integrity of digital images are often threatened by unauthorized manipulation that can obscure the source and validity of visual data. Weaknesses in conventional watermarking methods cause watermarks to be easily damaged when images undergo compression, noise, or geometric transformation. This study aims to analyze the effectiveness of a selective Least Significant Bit (LSB)-based watermarking method combined with a random key generation algorithm. This type of research is quantitative experimental research with a structured experimental evaluation design. The research subjects consisted of 90 test digital images covering variations in smooth, raw, and edge intensities. Data were collected through simulation-based computational testing, while the research instruments included the objective metrics PSNR, SSIM, MSE, and BER, which had been validated through replication reliability testing. Data analysis was performed using a descriptive statistical approach and a comparison of performance between image attack conditions to assess the stability of the watermarking results. The results show that watermark insertion in areas of smooth intensity provides the best resistance and invisibility. The conclusion of the study indicates that the intensity-based selective LSB watermarking method with random keys is effective for improving the security and reliability of digital image authentication. The implications of this research confirm that this approach has the potential to be applied to image authentication systems in devices with computational limitations, such as medical systems, digital archives, and IoT applications, to strengthen the traceability and integrity of visual data.]]>
