In-Situ High-Speed Computer Vision

← Back to all benchmarks

Date: 2023-12-05

Name: In-Situ High-Speed Computer Vision

Domain: High Energy Physics

Focus: Real-time image classification for in-situ plasma diagnostics

Task Types: Image Classification

Metrics: Accuracy, FPS

Models: CNN

AI/ML Motif: Classification

Resources

Benchmark: Visit

Results: Gemini LLM Deep Research, ChatGPT LLM

Keywords

plasma in-situ vision real-time ML

Citation

• Yumou Wei, Ryan F. Forelli, Chris Hansen, Jeffrey P. Levesque, Nhan Tran, Joshua C. Agar, Giuseppe Di Guglielmo, Michael E. Mauel, and Gerald A. Navratil. Low latency optical-based mode tracking with machine learning deployed on fpgas on a tokamak. 2024. URL: https://arxiv.org/abs/2312.00128, arXiv:2312.00128, doi:https://doi.org/10.1063/5.0190354.

@misc{wei2024lowlatencyopticalbasedmode,
  archiveprefix = {arXiv},
  author        = {Yumou Wei and Ryan F. Forelli and Chris Hansen and Jeffrey P. Levesque and Nhan Tran and Joshua C. Agar and Giuseppe Di Guglielmo and Michael E. Mauel and Gerald A. Navratil},
  doi           = {https://doi.org/10.1063/5.0190354},
  eprint        = {2312.00128},
  primaryclass  = {physics.plasm-ph},
  title         = {Low latency optical-based mode tracking with machine learning deployed on FPGAs on a tokamak},
  url           = {https://arxiv.org/abs/2312.00128},
  year          = {2024}
}

Ratings

CategoryRating

Software

1.00

No public implementation or containerized setup released

Specification

3.00

No standardized I/O, latency constraint, or complete framing

Dataset

0.00

Dataset not provided or described in any formal way

Metrics

2.00

Throughput and accuracy mentioned, but not defined or benchmarked

Reference Solution

1.00

Prototype CNNs described; no code, baseline, or training details available

Documentation

2.00

Some insight via papers, but no working repo, setup, or replication path

Average rating: 1.50/5

Radar plot

Edit: edit this entry