We have made an Android application that can detect hyperthermia in the diabetic foot as shown in this 25-second video.
The acquisition system consists of a smartphone and a FlirOne Pro thermal camera clipped onto the phone. The patient will be positioned on a couch and a thermal photo is taken from the feet. An algorithm based on artificial intelligence isolates the arches of the foot in the thermal image. The point-to-point temperature difference between the left and right foot is then displayed. Red areas show the locations where that difference is highest.
The thermal difference greater than 2.2 ° C is recognised as hyperthermia that is deemed to be a precursor of an ulcer.
For the thermal image of the patients analysed here, shows an average difference is 0.8 ° C and that the hottest areas (at the right foot heel area and at the left big toe region ) which are less than 2.2 ° C.
This easy-to-use and very fast system will be offered for sale for hospitals and medical practices to better prevent ulcers in the diabetic foot. This can contribute to better care for diabetic patients.
Below you will find the scientific publications of this work.
[1] D. Bouallal, H. Douzi, R. Harba, Diabetic foot thermal image segmentation using Double Encoder-ResUnet (DE-ResUnet), Journal of Medical Engineering & Technology, Vol. 46:5, pp. 378-392, 2022.
[2] D. Bouallal, A. Bougrine, R. Harba, R. Canals, H. Douzi, L. Vilcahuaman, H. Arbanil, STANDUP database of plantar foot thermal and RGB images for early ulcer detection, Open Research Europe, https://doi.org/10.12688/openreseurope. 14706.1, 2022.
[3] D. Bouallal, H. Douzi, R. Harba, Registration Methods for Thermal Images of Diabetic Foot Monitoring: A Comparative Study, International Journal of Advanced Computer Science and Applications, doi 10.14569/IJACSA.2022.0130670, 2022.
[4] A. Bougrine, R. Harba, R. canals, R . Ledee, M. Jabloun, A. Villeneuve, Segmentation of Plantar Foot Thermal Images Using Prior Information, Sensors 2022, 22(10), 3835; https://doi.org/10.3390/s22103835
[5] A. Bougrine, R. Harba, R. Canals, R. Ledée and M. Jabloun, A comparison of active contours prior shape segmentation methods: application to diabetic plantar foot thermal images, ICDIPV, doi:10.5121/csit.2019.90404, 2019.
[6] A. Bougrine, R. Harba, R. Canals, R. Ledée and M. Jabloun, On the segmentation of plantar foot thermal images with Deep Learning, EUSIPCO, doi:10.23919/EUSIPCO.2019.8902691, 2019.
[7] A. Bougrine, R. Ledee, R. Canals, R. Harba, M. Jabloun, Segmentation d’images thermiques de la voûte plantaire par Deep Learning, GRETSI, 2019.
[8] A. Bougrine, R. Harba, R. Canals, R. Ledee, and M. Jabloun, A mobile, reliable and user-friendly technology for processing plantar foot infrared thermal images applied to the early prevention of foot ulcers, DiabeticFoot-Europe, 2019.
[9] D. Bouallal, A. Bougrine, H. Douzi, R. Harba, and R. Canals, Segmentation of plantar foot thermal images: Application to diabetic foot diagnosis, International Conference on Systems, Signals, and Image Processing, vol. 2020-July, pp. 116–121, 2020.
[10] A. Bougrine, R. Canals, R. Harba, K. Jin, Smartphone hyperthermia detection in plantar foot thermal images with Deep Learning, ISBI, 2021.
[11] D. Bouallal, H. Douzi, R. Harba, Segmentation of plantar foot thermal images: application to diabetic foot diagnosis, International Conference on Systems, Signals and Image Processing, pp. 116-121, 2020.
[12] A. Aferhane, D. Bouallal, H. Douzi, R. Harba, Affine registration of plantar foot thermal images with Deep Learning: application to diabetic foot diagnosis, The International Conference on Intelligent Systems and Smart Technologies, 2023.
[13] A. Bougrine Analyse d’images thermiques de la voute plantaire : application au diagnostic du pied diabétique. PhD defence the 8 October 2020, Université of Orléans.
[14] D. Bouallal Analyse et traitement des images thermiques pour le contrôle et l'évaluation du pied diabétique par apprentissage profond. PhD defence the 18 March 2023, university of Agadir.