Yolo-MediSpecs: Deep Yolo Network based Medicine Expiry Date Detection assisted Specs for Visually Impaired People

Blind people and people with vision impairments require assistance to minimize their risk of colliding with external objects. Every aspect of human existence is now impacted by technology, and people use new gadgets frequently. A lack of specialized knowledge or real-time dynamics make dates less reliable. A novel deep learning-based Yolo-MediSpecs model has been developed to assist visually impaired (VI) people in managing medication expiration dates. Initially, medication packets are photographed by sensors in the specifications, such as the camera, Raspberry Pi, TOCR, and headphones. Images are processed using Multi-Scale Retinex (MSR) and background removal. The trained model is then kept from overfitting by using improved pre-processed images. The detection decoder filter integrated Yolo (DDF-YOLO) model is a component of the proposed method. The DDF-YOLO algorithm is used to analyse a real-time image and identify expiration dates. In order to identify the day, month, and year (DMY) area independently, the filter output is sent to a decoupled attention network (DAN). A network of expiry detection is then accessed to determine whether the medication is expired. The VI person using hearing aids is also given information. In the experiment, the suggested system increased accuracy by 99.21%, demonstrating that it can accurately determine expiration dates. Based on an analysis of the proposed Yolo-MediSpecs against SmartMedBox, Med Glasses, EPSON BT-300, and third eye, the overall accuracy range is raised by 13.28%, 4.14%, 2.93%, and 10.05% respectively.