Ophthalmology stands out as a key domain for the advancement of Artificial Intelligence in the medical field because of its significant dependence on digital imaging methodologies. The abundance of visual data has fueled the creation of AI models that primarily focus on improving diagnostic accuracy. These models analyze medical images to identify eye diseases. These complex facilities could not be easily afforded in pre-diagnosis phase as examples, pre-scholar kids and populations in rural areas. In this work, we focus on the need for early detection of eye diseases, for kids, in order to identify the vulnerable cases and incent them to undertake their first medical examination. Within the framework of existing research, artificial intelligence and medical imagery are utilized for the purpose of diagnosing various pathologies. This particular study, however, seeks to leverage video streams obtained from smartphone cameras to conduct preliminary assessments for refractive errors and dry eye diseases, as well as identify susceptible individuals, particularly within the young population. In the present study, through the examination of a collection of smartphone videos of children eyes, we undergo the computation of the Eye Aspect Ratio (EAR), subsequently a deep learning pipeline is constructed with the specific aim of detecting two eye disorders, namely dry eye and refractive errors. the EAR dataset undergoes a transformation and augmentation process to become a spectrograms dataset. Finally, a Convolutional Neural Network (CNN) architecture is employed to analyze the spectrograms dataset, resulting in accuracy level of 82.5% for the identification of eye pathologies.