Android Prototype Recommends Eyeglasses by Detected Face Type

A mobile research prototype combines image processing, deep learning and internet-connected hardware to recommend eyeglass styles according to a person’s detected face type. The study, by Hasan Tezcan and Bekir Aksoy and published in Mobile Networks and Applications , describes an Android application that accepts a photograph taken with the phone’s camera or selected from its gallery, improves the image and sends it to another device operating online through the internet of things. The system then uses deep-learning methods to determine the face type and returns a list of eyeglass types associated with that classification.

The workflow links four components: image acquisition on an Android device, image enhancement, remote analysis and presentation of the recommendation. Custom Android software obtains the photograph through the device’s photo-taking feature or gallery. Image-processing techniques improve the picture before it is analyzed by a separate internet-connected device.

Following the face-type determination, the application displays the glasses types the system identifies as most suitable for the person. In this configuration, the phone provides the user interface while the computational analysis occurs through a networked device rather than entirely on the handset. Technically, the project draws on the increased computing capacity of mobile devices and the connectivity between mobile applications and other networked systems.

The authors describe deep learning as a subdiscipline of artificial intelligence with capabilities for complex calculation, abstract representation and pattern recognition. In the proposed system, those capabilities are used to analyze the enhanced portrait and assign a facial category, while the internet of things provides the connection between the Android application and the online analysis device. The system’s intended practical benefit is convenience: users can obtain a face-type classification and a corresponding list of frame options through a digital interface, potentially saving time and space.

Those savings are stated aims, not quantified results in the accessible material. The supplied abstract and preview also do not provide numerical measures of face-classification accuracy, recommendation quality, processing speed or reliability. They do not specify the neural-network architecture, training dataset, number of facial categories, criteria used to define each category, eyewear database, Android-device specifications or image-resolution requirements, preventing independent assessment of those components.

Because the available material consists of the abstract and a preview with numbered figures, the work is presented here as a research prototype and application concept. The accessible preview does not establish the size or composition of the dataset, the populations represented in it, the conditions under which images were captured or whether users participated in a quantified evaluation. It also does not explain whether face types are defined through facial proportions, geometric measurements or another approach, nor how the recommended glasses are mapped to each category.

No evidence in the accessible material shows whether the system improves appearance, fit, visual performance or clinical suitability. The reported development nevertheless demonstrates a concrete route from a mobile portrait to a personalized list of eyewear categories, using image enhancement and remote deep-learning analysis to connect face-type classification with a glasses recommendation.

Subject of Research: A mobile deep-learning prototype for face-type classification and eyeglass-style recommendations

Article Title: Implementation of a Mobile Supported Glasses Recommendation System According to Human Face Type Using Deep Learning and Internet of Things

Article References: Tezcan, H., & Aksoy, B. (2026). Implementation of a Mobile Supported Glasses Recommendation System According to Human Face Type Using Deep Learning and Internet of Things. Mobile Networks and Applications. https://doi.org/10.1007/s11036-026-02512-w

Image Credits: AI Generated

DOI: 10.1007/s11036-026-02512-w

Keywords: deep learning, Android app, face-type classification, eyeglass recommendations, image processing, internet of things, mobile computing, computer vision, pattern recognition, prototype system, Implementation, Mobile