Understanding issues on a professional level can be frustrating and hard for most people.
Though there are other applications that can analyze the target information, they may not work
well in a non-ideal environment. To address the problem, a natural scene information acquisition
and analysis model based on deep learning are proposed in this study. In this four-part model,
Connectionist Text Proposal Network text detection algorithm, projection-based text segmentation,
Capsule network text recognition, and text analysis methods are adapted to process the target information.
The result is that users can get intelligent solutions on a professional level through uploading
target information pictures in a non-ideal environment. The model has reduced deviations
due to problems such as special shooting conditions and shooting techniques by using deep learning
algorithms. In this study, the model is employed to identify the nutrient component table accurately
and find the relationship between nutrients and diseases and analyze the nutritional intake of
different users to achieve precision medicine, and finally provide dietary advice for users.
Keywords: Scene Text Detection, Scene Text Recognition, Deep Learning, Precision Medicine, Nutritional Requirements, Healthy Diet
Rights & PermissionsPrintExport