Deep learning in dental diagnostics: Caries detection through smartphone photographs – A systematic review

Tooth decay is a common problem worldwide and detecting it early is crucial in preventing serious complications at a later stage. However, many people, due to socioeconomic factors, geographical barriers, do not have easy access to dentists. This review looks at how deep learning, a subset of artificial intelligence (AI), can help detect caries using photographs captured with smartphones. Nowadays, smartphones are widely available and have good cameras that can take clear pictures of teeth. Deep learning models can analyze these pictures to identify cavities. The present study reviewed studies published between 2005 and 2025 taken from major research databases to evaluate how well these technologies work for early cavity detection, especially for people with limited dental care. The findings show that deep learning models using smartphone images can detect visible cavities with good accuracy. Methods such as improving image quality and combining different deep learning techniques made the detection better. This approach is low-cost and easy to use, which makes it ideal for basic dental screenings in low-income or hard-to-reach areas. However, detecting very early-stage cavities is still challenging with this approach. Factors such as saliva, lighting, and camera angles can lower the quality of the pictures and affect the performance of these AI models. In addition, these models need large and varied collections of tooth images to train the models properly, but gathering these can be expensive and challenging. Using deep learning with images captured through a smartphone offers a promising and accessible way to screen for tooth decay. More research is needed to improve the detection of early cavities and to build larger, more diverse image databases to help train these models better. This technology could make dental care easier to reach many people around the world.