Longitudinal Infrared Image Registration Without Adhesive Markers for Breast Cancer Chemotherapy Patients

Lee, Chia-Yen; Chen, Chung–Ming; Wang, Hao-Jen; Lai, J.L.; Lee, Chi-En; Lin, Fangzhen; Chung, Ming-Jui; Chang, Yeun‐Chung

doi:10.1166/jmihi.2018.2314

“…After the corresponding pairs were found well, the relationship between the two IR images was established by thin plate spline (TPS), which was employed in image transformation to simulate the surface deformation of nonrigid objects such as the human breast (Lee et al 2018).…”

Section: Methodsmentioning

confidence: 99%

Automated marker-free longitudinal infrared breast image registration by GA-PSO

Lin,

Lee,

Chen

et al. 2023

Phys. Med. Biol.

Self Cite

0

View full text Add to dashboard Cite

The automated marker-free longitudinal Infrared (IR) breast image registration overcomes several challenges like no anatomic fiducial markers on the body surface, blurry boundaries, heat pattern variation by environmental and physiological factors, nonrigid deformation etc., has the ability of quantitative pixel-wise analysis with the heat energy and patterns change in a time course study. To achieve the goal, Scale invariant feature transform (SIFT), Harris corner and Hessian matrix were employed to generate the feature points as anatomic fiducial markers, and hybrid genetic algorithm and particle swarm optimization (GA-PSO) minimizing the matching errors was used to find the appropriate corresponding pairs between the 1st IR image and the nth IR image. Moreover, the mechanism of the IR spectrogram hardware system has a high level of reproducibility. The performance of proposed longitudinal image registration system was evaluated by the simulated experiments and the clinical trial. In the simulated experiments, the mean difference of our system is 1.64 mm, which increases 57.58% accuracy than manual determination and makes 17.4% improvement than the previous study. In the clinical trial, 80 patients were captured several times of IR breast images during chemotherapy. Most of them were well aligned in the spatiotemporal domain. In the few cases with evident heat pattern dissipation and spatial deviation, it still provided the reliable comparison of vascular variation. Therefore, the proposed system is accurate and robust, which could be considered as a reliable tool for longitudinal approaches of breast cancer diagnosis.

show abstract

“…4) The image deformation part: The outlier pairs were eliminated by random sample consensus (RANSAC). After the corresponding pairs were found well, the relationship between the two IR images was established by thin plate spline (TPS), which was employed in image transformation to simulate the surface deformation of nonrigid objects such as the human breast [21].…”

Section: Methodsmentioning

confidence: 99%

“…2) The feature point detection part: The heat pattern and the branch points of the skeletons of the heat pattern on the breast surface as the initial fiducial points were implemented by Harris corner detector algorithm and Hessian matrix to find corner points and cross-points of the middle lines in the vessel heat pattern, respectively [21,22]. Moreover, SIFT has better performance than SURF in resisting interference of changes in scale and rotation [52].…”

Section: Methodsmentioning

confidence: 99%

Automated Marker-Free Longitudinal Infrared Breast Image Registration by GA-PSO

Lee

¹

,

Chen

²

,

Lin

³

et al. 2022

Preprint

Self Cite

0

View full text Add to dashboard Cite

The automated marker-free longitudinal Infrared (IR) breast image registration overcomes several challenges like no anatomic fiducial markers on the body surface, blurry boundaries, heat pattern variation by environmental and physiological factors, nonrigid deformation etc., has the ability of quantitative pixel-wise analysis with the heat energy and patterns change in a time course study. To achieve the goal, Scale invariant feature transform (SIFT), Harris corner and Hessian matrix were employed to generate the feature points as anatomic fiducial markers, and hybrid genetic algorithm and particle swarm optimization (GA-PSO) minimizing the matching errors was used to find the appropriate corresponding pairs between the 1st IR image and the nth IR image. Moreover, the mechanism of the IR spectrogram hardware system has a high level of reproducibility. The performance of proposed longitudinal image registration system was evaluated by the simulated experiments and the clinical trial. In the simulated experiments, the mean difference of our system is 1.64 mm, which increases 57.58% accuracy than manual determination and makes 17.4% improvement than the previous study. In the clinical trial, 80 patients were captured several times of IR breast images during chemotherapy. Most of them were well aligned in the spatiotemporal domain. In the few cases with evident heat pattern dissipation and spatial deviation, it still provided the reliable comparison of vascular variation. Therefore, the proposed system is accurate and robust, which could be considered as a reliable tool for longitudinal approaches of breast cancer diagnosis.

show abstract

“…Since the project of this study is human breast, which is non-linearity and unstable. Accordingly, this study adopts the non-linear thin plate spline (TPS) 40 41 which belongs to the curved transformation model.…”

mentioning

confidence: 99%

Automatic Marker-free Longitudinal Infrared Image Registration by Shape Context Based Matching and Competitive Winner-guided Optimal Corresponding

Lee

¹

,

Wang

²

,

Lai

³

et al. 2017

Sci Rep

Self Cite

View full text Add to dashboard Cite

Long-term comparisons of infrared image can facilitate the assessment of breast cancer tissue growth and early tumor detection, in which longitudinal infrared image registration is a necessary step. However, it is hard to keep markers attached on a body surface for weeks, and rather difficult to detect anatomic fiducial markers and match them in the infrared image during registration process. The proposed study, automatic longitudinal infrared registration algorithm, develops an automatic vascular intersection detection method and establishes feature descriptors by shape context to achieve robust matching, as well as to obtain control points for the deformation model. In addition, competitive winner-guided mechanism is developed for optimal corresponding. The proposed algorithm is evaluated in two ways. Results show that the algorithm can quickly lead to accurate image registration and that the effectiveness is superior to manual registration with a mean error being 0.91 pixels. These findings demonstrate that the proposed registration algorithm is reasonably accurate and provide a novel method of extracting a greater amount of useful data from infrared images.

show abstract

Longitudinal Infrared Image Registration Without Adhesive Markers for Breast Cancer Chemotherapy Patients

Cited by 3 publications

References 0 publications

Automated marker-free longitudinal infrared breast image registration by GA-PSO

Automated marker-free longitudinal infrared breast image registration by GA-PSO

Automated Marker-Free Longitudinal Infrared Breast Image Registration by GA-PSO

Automatic Marker-free Longitudinal Infrared Image Registration by Shape Context Based Matching and Competitive Winner-guided Optimal Corresponding

Contact Info

Product

Resources

About