MULTI-VIEW PROTOTYPE

A prototype has been developed and patented [1] that allows imaging at both macro and micro scales. It has been validated with a healthspan experiment for classification of the different movement patterns of neurodegenerative disease models of Hungtinton [2].

New tracking and focusing [3] techniques have been developed and integrated into the multi-view prototype.

Using the micro-images captured by the system, a method for solving the re-identification problem based on neural networks has been proposed and promising results have been obtained.

To ensure that the identification network learns the internal characteristics of the nematode, it is necessary to position the nematodes straight

This transformation can be performed using spatial transformation networks (STNs). In this project, a new strategy that improves the convergence of these networks was studied [4].

References

[1] Sánchez-Salmerón, A. J., Puchalt, J. C., Gonzalez-Rojo, J. F. y Ricolfe C.  “SISTEMA Y PROCEDIMIENTO AUTOMATIZADO PARA LA MONITORIZACIÓN A NIVEL MICROESCALA DE LOS C. ELEGANS CONTENIDOS EN UNA PLACA PETRI CON MEDIO TRANSPARENTE, REGISTRO DE MONITORIZADO OBTENIDO Y USO DEL MISMO”. Número de solicitud: 202130492 (2021) 

[2] Puchalt, J. C., Gonzalez-Rojo, J. F., Gómez, A., Vázquez, R., & Sánchez-Salmerón, A. J. (2022). Multiview motion tracking based on a cartesian robot to monitor Caenorhabditis elegans in standard Petri dishes. Scientific Reports, 12(1), 1-11.

[3] Frangione, A., Salmerón, A. S., Modica, F., & Percoco, G. (2019). Multi-step approach for automated scaling of photogrammetric micro-measurements. The International Journal of Advanced Manufacturing Technology, 102(1-4), 747-757.

[4] Navarro Moya, F., Puchalt, J. C., Layana Castro, P. E., García Garví, A., & Sánchez-Salmerón, A. J. (2022). A new training strategy for spatial transform networks (STN’s). Neural Computing and Applications.