Implementation of artificial intelligence and non-contact infrared thermography for prediction and personalized automatic identification of different stages of cellulite
Background Cellulite is a common physiological condition of dermis, epidermis, and subcutaneous tissues experienced by 85 to 98% of the post-pubertal females in developed countries. Infrared (IR) thermography combined with artificial intelligence (AI)-based automated image processing can detect both...
Gespeichert in:
Veröffentlicht in: | The EPMA journal 2020-03, Vol.11 (1), p.17-29 |
---|---|
Hauptverfasser: | , , , , , , , |
Format: | Artikel |
Sprache: | eng |
Schlagworte: | |
Online-Zugang: | Volltext |
Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
Zusammenfassung: | Background
Cellulite is a common physiological condition of dermis, epidermis, and subcutaneous tissues experienced by 85 to 98% of the post-pubertal females in developed countries. Infrared (IR) thermography combined with artificial intelligence (AI)-based automated image processing can detect both early and advanced cellulite stages and open up the possibility of reliable diagnosis. Although the cellulite lesions may have various levels of severity, the quality of life of every woman, both in the physical and emotional sphere, is always an individual concern and therefore requires patient-oriented approach.
Objectives
The purpose of this work was to elaborate an objective, fast, and cost-effective method for automatic identification of different stages of cellulite based on IR imaging that may be used for prescreening and personalization of the therapy.
Materials and methods
In this study, we use custom-developed image preprocessing algorithms to automatically select cellulite regions and combine a total of 9 feature extraction methods with 9 different classification algorithms to determine the efficacy of cellulite stage recognition based on thermographic images taken from 212 female volunteers aged between 19 and 22.
Results
A combination of histogram of oriented gradients (HOG) and artificial neural network (ANN) enables determination of all stages of cellulite with an average accuracy higher than 80%. For primary stages of cellulite, the average accuracy achieved was more than 90%.
Conclusions
The implementation of computer-aided, automatic identification of cellulite severity using infrared imaging is feasible for reliable diagnosis. Such a combination can be used for early diagnosis, as well as monitoring of cellulite progress or therapeutic outcomes in an objective way. IR thermography coupled to AI sets the vision towards their use as an effective tool for complex assessment of cellulite pathogenesis and stratification, which are critical in the implementation of IR thermographic imaging in predictive, preventive, and personalized medicine (PPPM). |
---|---|
ISSN: | 1878-5077 1878-5085 |
DOI: | 10.1007/s13167-020-00199-x |