| CPC G16H 50/20 (2018.01) [G06T 7/0012 (2013.01); G06V 10/82 (2022.01); G16H 30/20 (2018.01); G16H 50/70 (2018.01); G06T 2207/20081 (2013.01)] | 6 Claims |
|
1. An artificial intelligence-based scalp image diagnostic analysis system using big data, the system comprising:
a main processor (3) configured to: receive, from a diagnostician, information about a customer's history taken by the diagnostician by asking about the customer's history, and a scalp image obtained by any one of a scalp diagnosis device and a terminal, through API (RESTful) (2) as a cloud service; conduct a diagnosis by a self-diagnosis algorithm with respect to the received history-taking information; and transmit the received scalp image to an artificial-intelligence processor (5), for performing a scalp diagnosis;
the artificial-intelligence processor (5) configured to perform an artificial intelligence (AI) analysis to label the scalp image received from the main processor (3) with all or some of diagnosis items by use of data accumulated in database (4);
wherein the artificial-intelligence processor (5) learns about the scalp by the artificial intelligence (AI) analysis using information of big data and collect learning data as a deep learning stage, labels the collected learning data, conducts learning and verification to label the collected data with learning data and test data, and derives an inference model (CNN: Convolutional Neural Network),
wherein the deep learning conducts scalp labelling (CNN: object recognition) through retraining by use of TensorFlow and an Inception V3 model, whereby the scalp is labelled with all or some of the diagnosis items,
a scalp diagnosis AI algorithm (6) configured to: receive, from the artificial-intelligence processor (5), information labeled with all or some of the diagnosis items; conduct a specific precision diagnosis by performing learning and interpretation by a deep learning algorithm, and derive a final diagnosis result; and
the database (4) accumulating therein scalp measurement, diagnosis, and recommendation data, which are provided to the main processor, thereby enabling self-scalp a diagnosis and recommendation service to be performed,
wherein the artificial-intelligence processor counts multiple numbers of hair follicle groups and multiple number of follicles within each group based on a microscopic image of a sample from a human scalp, and
wherein the diagnosis items include the scalp types which are dry, sensitive, inflammatory, with hair loss, good, oily, scurfy, and seborrhoeic.
|