International Conference on Science and Spirituality for Global Peace and Harmony

IAPIC-2025, Hyderabad, Telangana State, India (April 9-12, 2025)

A REVIEW ON ADVANCEMENTS IN
BIOMEDICAL IMAGE PROCESSING: 2015–2025

Prashanthi Govindarajan
Consultant
A3IT Tech Inc.
Toronto, Canada

Abstract
In the last decade, biomedical image processing has made substantial progress. It could be largely attributed to
the development of deep learning and artificial intelligence (AI) and innovative ways it is incorporated to
biomedical image processing. This paper examines major advancements in the field from 2015 to 2025, with a
broad focus on three areas: deep tomographic reconstruction, AI -powered diagnostic tools, and cutting-edge
methods for medical image segmentation. These innovations have significantly enhanced the precision and
effectiveness of medical imaging, offering promising possibilities for more individualized and prompt healthcare.
The objective of this paper is to review these developments, identify ongoing challenges, and explore the potential
for AI technologies to continue transforming the landscape of biomedical imaging.

1. Introduction
Our human body is a complex, well organized system that consists of several other component systems like
cardiovascular system, nervous system to name a few. And each of the major systems in our body is made up of
subsystems that help carry various physiological processes. All these systems work in cohesion to maintain
homeostasis, loosely translated to what is known as equilibrium. The input s for the systems in our body could be
neurotransmitters or information, physical material and corresponding outputs may be mechanical signals,
electrical or biochemical in nature. Any ailment in the body causes alterations to these signals and subsequently
affects the well being of the body. Observing and analysing these systems, thus could help understand the system
better and provide better pathways to treatment of the ailment [1].
Biomedical image processing plays a vital role in modern healthcare, enabling precise diagnostics and improving
the planning of treatments. Over the past decade, from 2015 to 2025, this field has experienced significant
advancements, primarily driven by the integration of deep learning and artificial intelligence (AI) technologies.
These innovations have revolutionized how medical images are analyzed, leading to better detection of diseases,
more accurate interpretation of imaging data, and enhanced decisio n-making processes. AI algorithms,
particularly those based on deep learning, have shown remarkable performance in tasks such as image
segmentation, feature extraction, and pattern recognition. As a result, healthcare providers are now able to identify
abnormalities with greater accuracy and speed, ultimately improving patient outcomes. Additionally, AI -driven
advancements in imaging techniques have facilitated the development of more personalized treatment plans,
tailored to individual patient needs. This ongoing transformation marks a critical shift toward more efficient,
automated, and effective healthcare solutions [2-4].
In this paper, a review of these developments has been presented alongside the ongoing challenges, and the
potential for AI technologies that to continue transform the landscape of biomedical imaging has been discussed.

2. Deep Tomographic Reconstruction

The novel method of Deep Tomographic Reconstruction leverages deep learning techniques to enhance image
quality across various medical imaging modalities such as MRI, CT, SPECT, PET, Optical Imaging and
Ultrasound Imaging, and it addresses challenges like measurement noise, data sparsity, image artifacts, and
computational inefficiency. Significant advancements include:
• Computed Tomography (CT): In Computed Tomography (CT), deep learning models have successfully
minimized radiation exposure while preserving image quality. One example is the use of residual
encoder-decoder convolutional neural networks for low-dose CT imaging, which significantly reduces
the radiation dose without compromising the quality of the images.
• Magnetic Resonance Imaging (MRI): In Magnetic Resonance Imaging (MRI), deep learning has
enhanced acquisition speeds, referred to as fast MRI, and helped reduce motion artifacts. Methods like
domain-transform manifold learning have been applied to reconstruct high -quality images from under-
sampled data, leading to a notable decrease in scan durations.
• Positron Emission Tomography (PET) & Single Photon Emission CT (SPECT): In Positron Emission
Tomography (PET) and Single Photon Emission CT (SPECT), deep learning models have enhanced low-
dose imaging and corrected motion artifacts. The use of multimodal networks that incorporate MRI data

297
 International Conference on Science and Spirituality for Global Peace and Harmony
IAPIC-2025, Hyderabad, Telangana State, India (April 9-12, 2025)

has improved PET image quality by utilizing anatomical details. For example, deep neural networks
optimized for denoising PET images have led to significant improvements in image clarity.
These advancements highlight the profound effect of deep learning on tomographic image reconstruction,
resulting in improved diagnostic accuracy and more tailored patient care [5 -10].

3. AI-Enhanced Diagnostic Systems

There has been a significant improvement in detection of a disease and treatment planning with the integration of
AI into diagnostic systems. Remarkable ones include:
Cancer Diagnosis: AI models like the "Chief" foundation model, created by Harvard Medical School, have shown
impressive accuracy in identifying various types of cancer, evaluating treatments, and forecasting survival
outcomes. Trained on millions of unlabeled whole-slide tissue images, Chief has achieved a cancer detection
accuracy of up to 94%, surpassing previous AI approaches by as much as 36%.
Coronary Imaging: AI algorithms have significantly improved coronary imaging by providing real-time
evaluations of heart health, which supports preventive care and tailored treatment plans. Researchers at institutions
such as the University of Texas at San Antonio (UTSA) are working on AI-based methods to increase the precision
and effectiveness of coronary imaging.
There is a huge potential for AI in diagnostic systems by reducing human error and reducing the time involved
for assessment. Personalised treatment strategies for patients can be achieved at a rapid pace [11 -17].

4. Medical Image Segmentation Techniques

Recent advancements in medical image segmentation have significantly enhanced the precision and speed of
medical image analysis, leading to more accurate diagnoses and more effective treatment strategies. By leveraging
deep learning technologies, these advancements have minimized the reliance on manual efforts, automating the
process of segmenting complex anatomical structures in medical images. This has not only improved efficiency
but also reduced the risk of human error. A prominent example of this progress is the use of convolutional neural
networks (CNNs) in segmenting brain tumors in MRI scans. These AI -driven methods have demonstrated
remarkable accuracy, thus enabled earlier detection of tumors and provided clinicians with more detailed
information to guide treatment planning. Moreover, such advancements have contributed to faster turnaround
times for medical imaging, allowing healthcare providers to make quicker decisions and ultimately improving
patient outcomes [18-21].

5. Conclusion
From 2015 to 2025, significant progress has been made in biomedical image processing, largely due to the
incorporation of deep learning and AI technologies. These advancements have improved image reconstruction,
increased diagnostic precision, and refined segmentation methods, ultimately resulting in better patient outcomes
and more streamlined healthcare delivery. Medical image analysis has become far more accurate, efficient with
betterment in stability and scalability. In this paper, a review of the majo r advancements in biomedical image
processing in the decade 2015 to 2025 has been presented. The primary focus was on the areas of deep
tomographic reconstruction, AI-powered diagnostic tools, and cutting-edge methods for medical image
segmentation. These innovations have significantly enhanced the precision and effectiveness of medical imaging,
offering promising possibilities for more individualized and prompt healthcare.

References
1. R. M. Rangayyan and S. Krishnan, Biomedical Signal Analysis, 3rd ed. Hoboken, NJ: John Wiley &
Sons, 2024. ISBN 1119825873, 9781119825876.
2. H. T. Shen, X. Zhu, Z. Zhang, S.-H. Wang, Y. Chen, X. Xu, and J. Shao, “Heterogeneous data fusion for
predicting mild cognitive impairment conversion,” Information Fusion, vol. 66, pp. 54-63, 2021
3. Y. Zhu, M. Kim, X. Zhu, D. Kaufer, and G. Wu, “Long range early diagnosis of Alzheimer's disease
using longitudinal MR imaging data ,” Medical Image Analysis, vol. 67, p. 101825, 2021
4. Xiaoqing Liu, Kunlun Gao, Bo Liu, Chengwei Pan, Kongming Liang, Lifeng Yan, Jiechao Ma, Fujin
He, Shu Zhang, Siyuan Pan, et al. Advances in Deep Learning-Based Medical Image Analysis. Health
Data Sci. 2021; 2021:8786793. DOI:10.34133/2021/8786793
5. Sanuwani Dayarathna, Kh Tohidul Islam, Sergio Uribe, Guang Yang, Munawar Hayat, Zhaolin Chen,
Deep learning based synthesis of MRI, CT and PET: Review and analysis, Medical Image Analysis,
Volume 92, 2024, 103046, ISSN 1361-8415

298
 International Conference on Science and Spirituality for Global Peace and Harmony
IAPIC-2025, Hyderabad, Telangana State, India (April 9-12, 2025)

6. Zhang, H., & Qie, Y. (2023). Applying Deep Learning to Medical Imaging: A Review. Applied
Sciences, 13(18), 10521. https://doi.org/10.3390/app131810521
7. M. H. Asghari and B. Jalali, “Edge detection in digital images using dispersive phase stretch ,”
International Journal of Biomedical Imaging, vol. 2015, Article ID 687819, pp. 1–6, 2015.
8. H. Chen et al., “Low-Dose CT With a Residual Encoder-Decoder Convolutional Neural Network,” IEEE
Transactions on Medical Imaging, vol. 36, no. 12, pp. 2524–2535, Dec. 2017.
9. B. Zhu et al., “Image reconstruction by domain-transform manifold learning,” Nature, vol. 555, pp. 487–
492, Mar. 2018.
10. K. Gong et al., “PET Image Denoising Using a Deep Neural Network Through Fine Tuning,” IEEE
Transactions on Radiation and Plasma Medical Sciences, vol. 3, no. 2, pp. 153–161, Mar. 2019.
11. “AI breakthrough raises hopes for better cancer diagnosis,” Financial Times, Sep. 2024. [Online].
Available: https://www.ft.com/content/0a8f2c61-77f4-43ce-87d2-a7b421bbda85
12. “5 ways AI is making health care better,” Axios, Dec. 2024. [Online]. Available:
https://www.axios.com/local/san-antonio/sponsored/5-ways-ai-is-making-health-care-better.
13. J. Yoo et al., “Deep Learning Diffuse Optical Tomography,” IEEE Transactions on Medical Imaging,
vol. 39, no. 12, pp. 4013–4023, Dec. 2020.
14. Nierengarten, M. B. (2025). New AI model shows promise for cancer diagnosis
Cancer, 131(3), e35715. https://doi.org/10.1002/cncr.35715
15. Tolu-Akinnawo OZ, Ezekwueme F, Omolayo O, Batheja S, Awoyemi T. Advancements in Artificial
Intelligence in Noninvasive Cardiac Imaging: A Comprehensive Review. Clin Cardiol. 2025
Jan;48(1):e70087. doi: 10.1002/clc.70087.
16. Reichert Christopher, "UTSA, UT Health San Antonio use AI to advance coronary heart disease
imaging," UTSA Today, Jan. 2024. [Online]. Available: https://www.utsa.edu/today/2024/01/story/utsa-
ut-health-san-antonio-use-AI-for-coronary-heart-disease.html.
17. Doolub G, Mamalakis M, Alabed S, Van der Geest RJ, Swift AJ, Rodrigues JCL, Garg P, Joshi NV,
Dastidar A. Artificial Intelligence as a Diagnostic Tool in Non-Invasive Imaging in the Assessment of
Coronary Artery Disease. Med Sci (Basel). 2023 Feb 24;11(1):20. doi: 10.3390/medsci11010020.
18. Khan MKH, Guo W, Liu J, Dong F, Li Z, Patterson TA, Hong H. Machine learning and deep learning
for brain tumor MRI image segmentation. Exp Biol Med (Maywood). 2023 Nov;248(21):1974 -1992.
doi: 10.1177/15353702231214259. Epub 2023 Dec 16. PMID: 38102956; PMCID: PMC10798183.
19. C. Qiao et al., “Rationalized deep learning super-resolution microscopy for sustained live imaging of
rapid subcellular processes,” Nature Biotechnology, vol. 41, pp. 407–416, Mar. 2023.
20. H. Deng et al., “Deep learning in photoacoustic imaging: a review,” Journal of Biomedical Optics, vol.
26, no. 4
21. Arkapravo Chattopadhyay, Mausumi Maitra, MRI-based brain tumour image detection using CNN based
deep learning method, Neuroscience Informatics, Volume 2, Issue 4, 2022, 100060, ISSN 2772 -5286,
https://doi.org/10.1016/j.neuri.2022.100060.

About the Author

Prashanthi Govindarajan is an Electronics Engineer with the best of technical and analytical
skills. She received her MSc. Electronic Engineering (Sept 2010 – Feb 2012) from
Staffordshire University with Distinction and Overall Average Grade Point - 12.17. Her project
title was Signal Sparsification with Discrete Rajan Transform (DRT): Principles, Properties
and Applications. This study aimed at introducing Generalized Discrete Rajan Transform
(DRT) as a potential sparsifying transform. The results point towards the success of DRT in
achieving higher degree of sparsity. Her Industrial Project Overview titled Speaker
Recognition Using Mel-Frequency Cepstrum Coefficients (MFCCS) aims to enable machines
to understand and identify a human speaker using feature extraction and feature matching
techniques. From a wide range of algorithms available for such processing, the algorithm using
Mel-Frequency Cepstrum Coefficients (MFCC) is considered for feature extraction. She has
published research papers (i) “A Unified Framework to Measure Sparsity for Uncorrelated
Random Signal”, Transactions on Emerging Telecommunications Technologie s (ETT), Wiley

299
 International Conference on Science and Spirituality for Global Peace and Harmony
IAPIC-2025, Hyderabad, Telangana State, India (April 9-12, 2025)

Publications, 2012; (ii) A Novel Approach To Extract Hidden Features of Ultrasound

Imageries, International Journal of Systemics, Cybernetics and Informatics, pp 66 -66, India,
July 2008; (iii) Design Of An Automatic Speaker Recognition System Using MFCC,
Vector Quantization And LBG Algorithm, International Journal of Systemics, Cybernetics
and Informatics, pp 34-42, India, October 09.

Mrs. Prashanthi Govindarajan has been rcognized as a ‘Distinguished Scientist’ by the steering committee
members of the international conference on science, spirituality for global peace and harmony (IAPIC 2025). This
citation was awarded to her in absentia and the PDF version sent to her. The organizing and steering committee
members congratulate him for her innovative contribution in the field of ‘Signal and Image Processing’.

300