End-to-end implementation of a convolutional neural network on a 3D-integrated image sensor with macropixel array - CEA - Commissariat à l’énergie atomique et aux énergies alternatives Access content directly
Journal Articles Sensors Year : 2023

End-to-end implementation of a convolutional neural network on a 3D-integrated image sensor with macropixel array

Abstract

3D-integrated focal-plane array image processor chips offer new opportunities to implement computer vision algorithms directly inside the sensor. Neural networks in particular can perform highly complex machine vision tasks, and therefore their efficient implementation in such imagers are of significant interest. However, studies with existing pixel-processor array chips have focused on the implementation of a subset of neural network components - notably convolutional kernels. In this work, we implement a continuous end-to-end pipeline for a convolutional neural network from the digitisation of incoming photons to the output prediction vector on a macropixel-processor (where a single processor acts on set of pixels) array chip. Our implementation performs inference between 265 and 310 frames per second directly inside of the sensor by exploiting the different levels of parallelism available.
Fichier principal
Vignette du fichier
sensors-23-01909-v2.pdf (2.95 Mo) Télécharger le fichier
Origin : Publisher files allowed on an open archive

Dates and versions

cea-03983243 , version 1 (10-02-2023)
cea-03983243 , version 2 (28-08-2023)

Licence

Attribution

Identifiers

Cite

Maria Lepecq, Thomas Dalgaty, William Fabre, Stéphane Chevobbe. End-to-end implementation of a convolutional neural network on a 3D-integrated image sensor with macropixel array. Sensors, 2023, 23 (4), pp.1909. ⟨10.3390/s23041909⟩. ⟨cea-03983243v2⟩
43 View
39 Download

Altmetric

Share

Gmail Facebook X LinkedIn More