Pediatric oral health detection using Swin transformer

Oral health exams in children are important for a variety of reasons, including their impact on a child's overall well-being and development. Traditional dental health detection methods still need to be improved, particularly in actual diag-nostic environments. Artificial intelligence (AI) has the potential to have a substantial impact on pediatric dentistry by providing innovative solutions to numerous difficulties in children's oral healthcare. By analyzing X-rays or intra-oral scans, AI can aid in the early detection of disorders like tooth decay, and developmental anomalies, when correction is most successful. AI can provide additional insight by analyzing X-rays and data, allowing dentists to make more accurate treatment plan decisions. Transformer is an instance of a deep neural network that was first employed for natural language processing. Its main foundation is the self-attention mechanism. In this study, we assess the swin transformer model's effectiveness in pediatric oral health detection. The data classes were enhanced as a pre-processing step using the fuzzy color approach, and the images that were structured with original X-rays were stacked. The proposed system achieved an accuracy of 85.34 percent. The result obtained implies that swin transformer has prospects in medical imaging and pre-training may be a useful strategy for enhancing its model's accuracy on oral health detection tasks.