Representational transfer from publicly available models is a promising technique for improving medical image classification, especially in long-tailed datasets with rare diseases. However, existing methods often overlook the frequency-dependent behavior of these models, thereby limiting their effectiveness in transferring representations and generalizations to rare diseases. In this paper, we propose FoPro-KD, a novel framework that leverages the power of frequency patterns learned from frozen pre-trained models to enhance their transferability and compression, presenting a few unique insights: 1) We demonstrate that leveraging representations from publicly available pre-trained models can substantially improve performance, specifically for rare classes, even when utilizing representations from a smaller pre-trained model. 2) We observe that pre-trained models exhibit frequency preferences, which we explore using our proposed Fourier Prompt Generator (FPG), allowing us to manipulate specific frequencies in the input image, enhancing the discriminative representational transfer. 3) By amplifying or diminishing these frequencies in the input image, we enable Effective Knowledge Distillation (EKD). EKD facilitates the transfer of knowledge from pre-trained models to smaller models. Through extensive experiments in long-tailed gastrointestinal image recognition and skin lesion classification, where rare diseases are prevalent, our FoPro-KD framework outperforms existing methods, enabling more accessible medical models for rare disease classification. Code is available at https://github.com/xmed-lab/FoPro-KD.

Breast magnetic resonance imaging (MRI) and X-ray mammography are two image modalities widely used for early detection and diagnosis of breast diseases in women. The combination of these modalities, traditionally done using intensity-based registration algorithms, leads to a more accurate diagnosis and treatment, due to the capability of co-localizing lesions and susceptibles areas between the two image modalities. In this work, we present the first attempt to register breast MRI and X-ray mammographic images using intensity gradients as the similarity measure. Specifically, a patient-specific biomechanical model of the breast, extracted from the MRI image, is used to mimic the mammographic acquisition. The intensity gradients of the glandular tissue are directly projected from the 3D MRI volume to the 2D mammographic space, and two different gradient-based metrics are tested to lead the registration, the normalized cross-correlation of the scalar gradient values and the gradient correlation of the vectoral gradients. We compare these two approaches to an intensity-based algorithm, where the MRI volume is transformed to a synthetic computed tomography (pseudo-CT) image using the partial volume effect obtained by the glandular tissue segmentation performed by means of an Expectation-Maximization algorithm. This allows us to obtain the digitally reconstructed radiographies by a direct intensity projection. The best results are obtained using the scalar gradient approach along with a transversal isotropic material model, obtaining a target registration error (TRE), in millimeters, of 5.65 ± 2.76 for CC- and of 7.83 ± 3.04 for MLO-mammograms, while the TRE is 7.33 ± 3.62 in the 3D MRI. We also evaluate the effect of the glandularity of the breast as well as the landmark position on the TRE, obtaining moderated correlation values (0.65 and 0.77 respectively), concluding that these aspects need to be considered to increase the accuracy in further approaches.

Forced spirometry testing is gradually becoming available across different healthcare tiers including primary care. It has been demonstrated in earlier work that commercially available spirometers are not fully able to assure the quality of individual spirometry manoeuvres. Thus, a need to expand the availability of high-quality spirometry assessment beyond specialist pulmonary centres has arisen. In this paper, we propose a method to select and optimise a classifier using supervised learning techniques by learning from previously classified forced spirometry tests from a group of experts. Such a method is able to take into account the shape of the curve as an expert would during visual inspection. We evaluated the final classifier on a dataset put aside for evaluation yielding an area under the receiver operating characteristic curve of 0.88 and specificities of 0.91 and 0.86 for sensitivities of 0.60 and 0.82. Furthermore, other specificities and sensitivities along the receiver operating characteristic curve were close to the level of the experts when compared against each-other, and better than an earlier rules-based method assessed on the same dataset. We foresee key benefits in raising diagnostic quality, saving time, reducing cost, and also improving remote care and monitoring services for patients with chronic respiratory diseases in the future if a clinical decision support system with the encapsulated classifier is to be integrated into the work-flow of forced spirometry testing.

Breast lesion detection using ultrasound imaging is considered an important step of computer-aided diagnosis systems. Over the past decade, researchers have demonstrated the possibilities to automate the initial lesion detection. However, the lack of a common dataset impedes research when comparing the performance of such algorithms. This paper proposes the use of deep learning approaches for breast ultrasound lesion detection and investigates three different methods: a Patch-based LeNet, a U-Net, and a transfer learning approach with a pretrained FCN-AlexNet. Their performance is compared against four state-of-the-art lesion detection algorithms (i.e., Radial Gradient Index, Multifractal Filtering, Rule-based Region Ranking, and Deformable Part Models). In addition, this paper compares and contrasts two conventional ultrasound image datasets acquired from two different ultrasound systems. Dataset A comprises 306 (60 malignant and 246 benign) images and Dataset B comprises 163 (53 malignant and 110 benign) images. To overcome the lack of public datasets in this domain, Dataset B will be made available for research purposes. The results demonstrate an overall improvement by the deep learning approaches when assessed on both datasets in terms of True Positive Fraction, False Positives per image, and F-measure.

In this paper, we aim to produce a realistic 2-D projection of the breast parenchymal distribution from a 3-D breast magnetic resonance image (MRI). To evaluate the accuracy of our simulation, we compare our results with the local breast density (i.e., density map) obtained from the complementary full-field digital mammogram. To achieve this goal, we have developed a fully automatic framework, which registers MRI volumes to X-ray mammograms using a subject-specific biomechanical model of the breast. The optimization step modifies the position, orientation, and elastic parameters of the breast model to perform the alignment between the images. When the model reaches an optimal solution, the MRI glandular tissue is projected and compared with the one obtained from the corresponding mammograms. To reduce the loss of information during the ray-casting, we introduce a new approach that avoids resampling the MRI volume. In the results, we focus our efforts on evaluating the agreement of the distributions of glandular tissue, the degree of structural similarity, and the correlation between the real and synthetic density maps. Our approach obtained a high-structural agreement regardless the glandularity of the breast, whilst the similarity of the glandular tissue distributions and correlation between both images increase in denser breasts. Furthermore, the synthetic images show continuity with respect to large structures in the density maps.

Breast density measurement is an important aspect in breast cancer diagnosis as dense tissue has been related to the risk of breast cancer development. The purpose of this study is to develop a method to automatically compute breast density in breast MRI. The framework is a combination of image processing techniques to segment breast and fibroglandular tissue. Intra- and interpatient signal intensity variability is initially corrected. The breast is segmented by automatically detecting body-breast and air-breast surfaces. Subsequently, fibroglandular tissue is segmented in the breast area using expectation-maximization. A dataset of 50 cases with manual segmentations was used for evaluation. Dice similarity coefficient (DSC), total overlap, false negative fraction (FNF), and false positive fraction (FPF) are used to report similarity between automatic and manual segmentations. For breast segmentation, the proposed approach obtained DSC, total overlap, FNF, and FPF values of 0.94, 0.96, 0.04, and 0.07, respectively. For fibroglandular tissue segmentation, we obtained DSC, total overlap, FNF, and FPF values of 0.80, 0.85, 0.15, and 0.22, respectively. The method is relevant for researchers investigating breast density as a risk factor for breast cancer and all the described steps can be also applied in computer aided diagnosis systems.

A hybrid framework of probabilistic atlas and statistical shape and appearance model (SSAM) is proposed to achieve 3D prostate segmentation. An initial 3D segmentation of the prostate is obtained by registering the probabilistic atlas to the test dataset with deformable Demons registration. The initial results obtained are used to initialize multiple SSAMs corresponding to the apex, central and base regions of the prostate gland to incorporate local variabilities. Multiple mean parametric models of shape and appearance are derived from principal component analysis of prior shape and intensity information of the prostate from the training data. The parameters are then modified with the prior knowledge of the optimization space to achieve 2D segmentation. The 2D labels are registered to the 3D labels generated using probabilistic atlas to constrain the pose variation and generate valid 3D shapes. The proposed method achieves a mean Dice similarity coefficient value of 0.89±0.11 and mean Hausdorff distance of 3.05±2.25 mm when validated with 15 prostate volumes of a public dataset in a leave-one-out validation framework.

This paper presents a novel method to identify the 2D axial Magnetic Resonance (MR) slice from a pre-acquired MR prostate volume that closely corresponds to the 2D axial Transrectal Ultrasound (TRUS) slice obtained during prostate biopsy. The shape-context representations of the segmented prostate contours in both the imaging modalities are used to establish point correspondences using Bhattacharyya distance. Thereafter, Chi-square distance is used to find the prostate shape similarities between the MR slices and the TRUS slice. Normalized mutual information and correlation coefficient between the TRUS and MR slices are computed to find the information theoretic similarities between the TRUS-MR slices. The maximum of the weighted likelihood function of the afore-mentioned statistical similarity measures finally yields the MR slice that closely resembles the TRUS slice acquired during the biopsy procedure. The method is evaluated for 20 patient datasets and close matches with the ground truth are obtained for 16 cases.

Prostate volume estimation from segmented prostate contours in Trans Rectal Ultrasound (TRUS) images aids in diagnosis and treatment of prostate diseases, including prostate cancer. However, accurate, computationally efficient and automatic segmentation of the prostate in TRUS images is a challenging task owing to low Signal-To-Noise-Ratio (SNR), speckle noise, micro-calcifications and heterogeneous intensity distribution inside the prostate region. In this paper, we propose a probabilistic framework for propagation of a parametric model derived from Principal Component Analysis (PCA) of prior shape and posterior probability values to achieve the prostate segmentation. The proposed method achieves a mean Dice similarity coefficient value of 0.96±0.01, and a mean absolute distance value of 0.80±0.24 mm when validated with 24 images from 6 datasets in a leave-one-patient-out validation framework. Our proposed model is automatic, and performs accurate prostate segmentation in presence of intensity heterogeneity and imaging artifacts.

The detection of architectural distortions and abnormal structures in mammographic images can be based on the analysis of bilateral and temporal cases using image registration. This paper presents a quantitative evaluation of state-of-the art intensity based image registration methods applied to mammographic images. These methods range from a global and rigid transformation to local deformable paradigms using various metrics and multiresolution approaches. The aim of this study is to assess the suitability of these methods for mammographic image analysis. Evaluation using temporal cases based on quantitative analysis and a multiobserver study is presented which gives an indication of the accuracy and robustness of the different algorithms. Although previous studies suggested that local deformable methods were not suitable due to the generation of unrealistic distortions, in this work we show that local deformable paradigms (multiresolution B-Spline deformations) obtain the most accurate registration results.

The detection of architectural distortions and abnormal structures in mammographic images can be based on the analysis of bilateral and temporal cases using image registration. This work presents a quantitative evaluation of eight state-of-the art image registration methods applied to mammographic images. These methods range from a global and rigid transformation to local deformable paradigms using various metrics and multi-resolution approaches. The aim of this study is to assess the suitability of these methods for mammographic image analysis. Evaluation using temporal cases based on quantitative analysis gives an indication of the accuracy and robustness of the different algorithms. This work shows that local deformable paradigms (B-spline deformations) obtain the most accurate registration results.

Elastography, the imaging technique for estimating the elastic tissue properties, or more specifically elastic modulus imaging, are becoming important diagnosis tools in computer aided diagnosis system, specially focusing on ultrasound and MRI images. This technique still presents unsolved challenges in the analysis of deformations in sequences of images. The aim of this paper is twofold: to evaluate the applicability of the deformation field obtained by state of the art optical flow and image registration algorithms for elastic modulus imaging and to quantitatively evaluate two different methods for estimation of the elastic modulus distribution. Results show that optical-flow methods provide a slightly better reconstruction and that the reconstruction has been shown to be more accurate using the method proposed by Sumi et al.

We present in this paper a supervised approach for automatic detection of micro-calcifications. The system is based on learning the different morphology of the micro-calcifications using local features, which are extracted using a bank of filters. Afterwards, this set of features is used to train a pixel-based boosting classifier which at each round automatically selects the most salient one. Therefore, when a new mammogram is tested only the salient features are computed and used to classify each pixel of the mammogram as being part of a micro-calcification or actually being normal tissue. The experimental results shows the validity of our approach. Moreover, the robustness of our method is also demonstrated using a digitised database for the learning process and a different one for the testing, providing satisfactory results.

It has been shown that the accuracy of mammographic abnormality detection methods is strongly dependent on the breast tissue characteristics, where a dense breast drastically reduces detection sensitivity. In addition, breast tissue density is widely accepted to be an important risk indicator for the development of breast cancer. Here, we describe the development of an automatic breast tissue classification methodology, which can be summarized in a number of distinct steps: 1) the segmentation of the breast area into fatty versus dense mammographic tissue; 2) the extraction of morphological and texture features from the segmented breast areas; and 3) the use of a Bayesian combination of a number of classifiers. The evaluation, based on a large kappa = 0.81 and 0.67 for the two data sets) between automatic and expert-based Breast Imaging Reporting and Data System mammographic density assessment.

We present a framework for the registration and correspondence of magnetic resonance (MR) (three-dimensional data, 3D, data) and x-ray (two-dimensional data, 2D, data) mammographic images. The robustness of this work relies on the development of a novel method to establish nonlinear correspondence between modalities of different dimensionality, which also represent different physical tissue aspects. The correspondence is based on a 2D–2D matching process, which takes into account features from internal linear structures from both images and a measure of global similarity between modalities (Martí et al., International Journal of Pattern Recognition and Artificial Intelligence, vol. 16, no. 3, pp. 331–340, 2002). The 2D–3D correspondence relies on an intermediate step, which establishes registration between the 2D x-ray image and a projection of the 3D MR data. Initial quantitative and qualitative evaluation results, based on a small data set, are presented that show the validity of the developed approach.

A novel method to obtain point correspondence in pairs of images is presented. Our approach is based on automatically establishing correspondence between linear structures which appear in images using robust features such as orientation, width and curvature extracted from those structures. The extracted points can be used to register sets of images. The potential of the developed approach is demonstrated on mammographic images.

A method to correspond linear structures in mammographic images is pre-sented. Our approach is based on automatically establishing correspondence between linear structures which appear in images using robust features such as orientation, width and curvature extracted from those structures. The re-sulting correspondence is used to track linear structures and regions in mam-mographic images taken at different times. 1

We have developed a novel similarity measure to evaluate image correspondence. Our method is based on the mutual information between images. The main difference to other mutual information approaches is that we incorporate spatial information using grey-level co-occurrence matrices, leading to a more general measurement. We have used this technique to evaluate two registration algorithms (local affine transformation and thin plate splines) applied to a dataset of mammographic images.