# On-Sensor Convolutional Neural Networks with Early-Exits

arXiv: https://arxiv.org/abs/2503.16939 · HTML: https://mlsanity.com/paper/2503.16939/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- Builds on 7 of its 15 references that are in the map.
- First posted on arXiv 21 Mar 2025.

## Builds on (most cited first, 7 of 7)

- [TensorFlow Lite Micro: Embedded Machine Learning on TinyML Systems](https://mlsanity.com/paper/2010.08678/) (2020)
- [Benchmarking TinyML Systems: Challenges and Direction](https://mlsanity.com/paper/2003.04821/) (2020)
- [Why should we add early exits to neural networks?](https://mlsanity.com/paper/2004.12814/) (2020)
- [NACHOS: Neural Architecture Search for Hardware Constrained Early Exit Neural Networks](https://mlsanity.com/paper/2401.13330/) (2024)
- [Machine Learning In-Sensors: Computation-enabled Intelligent Sensors For Next Generation of IoT](https://mlsanity.com/paper/2407.21508/) (2024)
- [BrainSlug: Transparent Acceleration of Deep Learning Through Depth-First Parallelism](https://mlsanity.com/paper/1804.08378/) (2018)
- [Low-memory convolutional neural networks through incremental depth-first processing](https://mlsanity.com/paper/1804.10727/) (2018)

## Built on by (most cited first, 1 of 1)

- [An Algorithm for On-Sensor Agnostic Detection of Changes in Human Activity for Ultra-Low-Power Applications](https://mlsanity.com/paper/2605.00870/) (2026)

## Abstract (arXiv)

Tiny Machine Learning (TinyML) is a novel research field aiming at integrating Machine Learning (ML) within embedded devices with limited memory, computation, and energy. Recently, a new branch of TinyML has emerged, focusing on integrating ML directly into the sensors to further reduce the power consumption of embedded devices. Interestingly, despite their state-of-the-art performance in many tasks, none of the current solutions in the literature aims to optimize the implementation of Convolutional Neural Networks (CNNs) operating directly into sensors. In this paper, we introduce for the first time in the literature the optimized design and implementation of Depth-First CNNs operating on the Intelligent Sensor Processing Unit (ISPU) within an Inertial Measurement Unit (IMU) by STMicroelectronics. Our approach partitions the CNN between the ISPU and the microcontroller (MCU) and employs an Early-Exit mechanism to stop the computations on the IMU when enough confidence about the results is achieved, hence significantly reducing power consumption. When using a NUCLEO-F411RE board, this solution achieved an average current consumption of 4.8 mA, marking an 11% reduction compared to the regular inference pipeline on the MCU, while having equal accuracy.