Alberto Ancilotto

[Elisabetta Farella @EfarellaFarella]

Overall 46mJ/frame and 6fps @ 280mW on a #TinyML device with RISC V dual core at 400MHz and a convolutional accelerator. 0.98 M Param and 0.064 G MACC. Thanks to the Phinets family in its new GAN version! @aaancilotto @fpaissan_

[Elisabetta Farella @EfarellaFarella]

Our demo at TinyML EMEA IF. Alberto face is swapped with the one of Nicolas Cage in real time for demonstrating video anonymization at the very edge as required for @MARVELprojecteu @TinymlF @aaancilotto @E3DA_FBK @FBK_research #TinyML

[Elisabetta Farella @EfarellaFarella]

Meeting great colleagues that are also good friends at TinyML EMEA 2022 Innovation Forum @TinymlF #TinyML @MicheleMagno5

[Elisabetta Farella @EfarellaFarella]

A great discussion during the break at EMEA Iinnovation forum 2022 with Adam Fuks from @NXP on quantization on #microcontrollers and several other topics! Adam, you are really a mine of knowledge! Thank you! @TinymlF #tinyml @E3DA_FBK

[Elisabetta Farella @EfarellaFarella]

Starting day 3 at #TinyML EMEA Innovation Forum 2022 with the keynote by Alberto San Giovanni Vincentelli @UCBerkeley. Starting well! @tinyMLTalks #NeuralNetworks #chips

[Elisabetta Farella @EfarellaFarella]

The main room

[E3DA Research Unit at DigiS center, FBK @E3DA_FBK]

The work we present is part of our research strategy, and it addresses the goals of the EU @MARVELprojecteu, where video anonymization addresses a smart city that protects the privacy of citizens while enabling LEA and PAs to benefit from multimodal data analytics

[Elisabetta Farella @EfarellaFarella]

We will present on Tuesday our work on GAN Implementation at the very edge @E3DA_FBK @FBK_research to which also @fpaissan_ contributed

[Elisabetta Farella @EfarellaFarella]

@aaancilotto and me at the #TinyML EMEA 2022 Forum... Room ready! Poster ready! We ready!