Bharath Sudharsan
ABSTRACT
TinyML: Tiny in size, big in impact! This paper presents TinyML-CAM pipeline for real-time and memory-efficient image recognition on IoT boards. TinyML-CAM can be used by developers to implement their customized tasks in ≈ 30 minutes, with minimal code configuration. We evaluated TinyML-CAM by using it to create a RandomForestClassifier (RF) based real-time image recognition system, which ran at 80 FPS and consumed only 1 kB of RAM on ESP32.
- B. Sudharsan, P. Patel, J. G. Breslin, and M. I. Ali. Enabling machine learning on the edge using sram conserving efficient neural networks execution approach. In Joint European Conference on Machine Learning and Knowledge Discovery in Databases, 2021.Google Scholar
Digital Library
- Eloquentarduino: https://github.com/eloquentarduino, 2022.Google Scholar
- Everywhereml: https://pypi.org/project/everywhereml/0.0.1/, 2022.Google Scholar
- B. Sudharsan, P. Yadav, J. G. Breslin, and M. I. Ali. An sram optimized approach for constant memory consumption and ultra-fast execution of ml classifiers on tinyml hardware. In IEEE International Conference on Services Computing (SCC), 2021.Google ScholarCross Ref
- C. Banbury, V. J. Reddi, et al. Mlperf tiny benchmark. Neural Information Processing Systems Track on Datasets and Benchmarks, 2021.Google Scholar
- B. Sudharsan. Training up to 50 class ml models on 3$ iot hardware via optimizing one-vs-one algorithm (student abstract). In Proceedings of the AAAI Conference on Artificial Intelligence.Google Scholar
Index Terms
- TinyML-CAM: 80 FPS image recognition in 1 kB RAM
Recommendations
TMM-TinyML: tensor memory mapping (TMM) method for tiny machine learning (TinyML)
MobiCom '22: Proceedings of the 28th Annual International Conference on Mobile Computing And NetworkingTinyML: Tiny in size, big in impact! In this paper, we present a Tensor Memory Mapping (TMM) method, which can accurately calculate the on-device execution memory consumed by a range of ML and TinyML models during execution on small central processing ...
Edge-assisted collaborative image recognition for augmented reality: demo abstract
SenSys '19: Proceedings of the 17th Conference on Embedded Networked Sensor SystemsMobile Augmented Reality (AR), which overlays digital information with real-world scenes surrounding a user, provides an enhanced mode of interaction with the ambient world. Contextual AR applications rely on image recognition to identify objects in the ...
A unified approach to coding and interpreting face images
ICCV '95: Proceedings of the Fifth International Conference on Computer VisionFace images are difficult to interpret because they are highly variable. Sources of variability include individual appearance, 3D pose, facial expression and lighting. We describe a compact parametrised model of facial appearance which takes into ...
Comments