In the period of extensive social media use, unconventional formats for news sharing often lack reliability, with misleading information disseminated through graphics emerging as a unique tactic. While the verification of the authenticity of news is a broad research domain, scholars have mainly focused on textual data. The misleading information detection model described in this paper uses neural sleuth-based networks to capture image data from the freely accessible Fakeddit dataset. The results are scrutinized using a confusion matrix, and a comprehensive evaluation of the model's performance metrics is conducted across six categories: “true,” “satire,” “false connection,” “imposter content,” “manipulated content,” and “misleading content.” This work contributes an inclusive analysis of the model's effectiveness as a problem posed by misleading information on social media.