Skip to main page content
U.S. flag

An official website of the United States government

Dot gov

The .gov means it’s official.
Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

Https

The site is secure.
The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2025 Feb 18;15(1):5846.
doi: 10.1038/s41598-025-88822-7.

Accelerating veterinary low field MRI acquisitions using the deep learning based denoising solution HawkAI

Jamil Nour Eddin  1 , Martin Blanchard  2 , Julien Guevar  3   4 , Valentina Curcio  5 , Hugo Dorez  6
Affiliations

Accelerating veterinary low field MRI acquisitions using the deep learning based denoising solution HawkAI

Jamil Nour Eddin et al. Sci Rep. .

Abstract

Magnetic resonance imaging (MRI) has changed veterinary diagnosis but its long-sequence time can be problematic, especially because animals need to be sedated during the exam. Unfortunately, shorter scan times implies a fall in overall image quality and diagnosis reliability. Therefore, we developed a Generative Adversarial Net-based denoising algorithm called HawkAI. In this study, a Standard-Of-Care (SOC) MRI-sequence and then a faster sequence were acquired and HawkAI was applied to the latter. Radiologists were then asked to qualitatively evaluate the two proposed images based on different factors using a Likert scale (from 1 being strong preference for HawkAI to 5 being strong preference for SOC). The aim was to prove that they had at least no preference between the two sequences in terms of Signal-to-Noise Ratio (SNR) and diagnosis. They slightly preferred HawkAI in terms of SNR (confidence interval (CI) being [1.924-2.176]), had no preference in terms of Artifacts Presence, Diagnosis Pertinence and Lesion Conspicuity (respective CIs of [2.933-3.113], [2.808-3.132] and [2.941-3.119]), and a very slight preference for SOC in terms of Spatial Resolution and Image Contrast (respective CIs of [3.153-3.453] and [3.072-3.348]). This shows the possibility to acquire images twice as fast without any major drawback compared to a longer acquisition.

Keywords: MRI; Machine learning; Veterinary imaging.

PubMed Disclaimer

Conflict of interest statement

Declarations. Competing interests: J.N. and M.B. are employees of HawkCell. V.C. was an employee of Hawkcell at the time of this research. H.D. is the founder and Chief Scientific Officer of HawkCell and owns stock in the company.

Figures

Fig. 1
Fig. 1
An example of a native image (left), a noisy image acquired more rapidly (middle), and the post-processed noisy image.
Fig. 2
Fig. 2
The Architecture of the HawkAI’s network, the generative adversarial network consisting of two main models, the discriminator (a classifier that distinguishes a generated image from a real image) and the generator (trained using the three loss functions).
Fig. 3
Fig. 3
Example of a dataset given to radiologists.
Fig. 4
Fig. 4
Example of a dataset given to radiologists.
Fig. 5
Fig. 5
Violin plot summary of the evaluation questionnaire (First image: HawkAI; Second Image: SOC).
Fig. 6
Fig. 6
Violin plot summary of the evaluation questionnaire (First image: HawkAI; Second Image: Noisy).
See this image and copyright information in PMC

References

    1. Gavin, P. Growth of clinical veterinary magnetic resonance imaging. Vet. Radiol. Ultrasound52, S2–S4. 10.1111/j.1740-8261.2010.01779.x (2011). - PubMed
    1. Thomson, C. E. et al. Magnetic resonance imaging’a general overview of principles and examples in veterinary neurodiagnosis. Vet. Radiol. Ultrasound34, 2–17. 10.1111/j.1740-8261.1993.tb01986.x (1993).
    1. Werpy, N. M. Magnetic resonance imaging of the equine patient: A comparison of high-and low-field systems. Clin. Tech. Equine Pract.6, 37–45. 10.1053/j.ctep.2006.11.004 (2007).
    1. Redpath, T. W. Signal-to-noise ratio in mri. Br. J. Radiol.71, 704–707. 10.1259/bjr.71.847.9771379 (1998). - PubMed
    1. Rudin, L. I. Nonlinear total variation based noise removal algorithms. Physica D60, 259–268. 10.1016/0167-2789(92)90242-F (1992).

LinkOut - more resources