Magnetic particle imaging (MPI) uses nonlinear response signals to noninvasively detect magnetic nanoparticles in space, and its quantitative properties hold promise for future precise quantitative treatments. In reconstruction, the system matrix based method necessitates suitable regularization terms, such as Tikhonov or non-negative fused lasso (NFL) regularization, to stabilize the solution. While NFL regularization offers clearer edge information than Tikhonov regularization, it carries a biased estimate of thel1penalty, leading to an underestimation of the reconstructed concentration and adversely affecting the quantitative properties. In this paper, a new nonconvex regularization method including min-max concave (MC) and total variation (TV) regularization is proposed. This method utilized MC penalty to provide nearly unbiased sparse constraints and adds the TV penalty to provide a uniform intensity distribution of images. By combining the alternating direction multiplication method (ADMM) and the two-step parameter selection method, a more accurate quantitative MPI reconstruction was realized. The performance of the proposed method was verified on the simulation data, the Open-MPI dataset, and measured data from a homemade MPI scanner. The results indicate that the proposed method achieves better image quality while maintaining the quantitative properties, thus overcoming the drawback of intensity underestimation by the NFL method while providing edge information. In particular, for the measured data, the proposed method reduced the relative error in the intensity of the reconstruction results from 28% to 8%.

Neoadjuvant chemotherapy (NAC) is the primary method to reduce the burden of tumor and metastasis; in the treatment of breast cancer, it may provide additional opportunities for breast-conserving surgery. Preoperative assessment of pathological complete response (PCR) to NAC is important for developing individualized treatment approaches and predicting patient prognosis. Compared to magnetic resonance imaging (MRI) and mammography, ultrasonography (US) has the advantages of simplicity, flexibility, and real-time imaging. Moreover, it does not require radiation and can provide multi-time acquisition of the tumor during NAC treatment. Recently, deep learning radiomics models based on multi-time-point US images for the prediction of NAC effectiveness have been proposed. To further improve the prediction performance, we carefully designed four supporting modules for our proposed dual-input transformer (DiT): isolated tokens-to-token patch embedding module, shared position embedding, time embedding, and weighted average pooling feature representation modules. The design of each module considers the characteristics of the US images at multiple time points. We validated our model on our retrospective US dataset composed of 484 cases from two centers whose consistency is not sufficiently high. Patients were allocated to training (n = 297), validation (n = 99), and external test (n = 88) sets. The results show that our model can achieve better performance than the Siamese CNN and the standard tokens-to-token vision transformer without using multi-time-point images. The ablation study also proved the effectiveness of each module designed for DiT.

Automatic semantic segmentation in 2D echocardiography is vital in clinical practice for assessing various cardiac functions and improving the diagnosis of cardiac diseases. However, two distinct problems have persisted in automatic segmentation in 2D echocardiography, namely the lack of an effective feature enhancement approach for contextual feature capture and lack of label coherence in category prediction for individual pixels. Therefore, in this study, we propose a deep learning model, called deep pyramid local attention neural network (PLANet), to improve the segmentation performance of automatic methods in 2D echocardiography. Specifically, we propose a pyramid local attention module to enhance features by capturing supporting information within compact and sparse neighboring contexts. We also propose a label coherence learning mechanism to promote prediction consistency for pixels and their neighbors by guiding the learning with explicit supervision signals. The proposed PLANet was extensively evaluated on the dataset of cardiac acquisitions for multi-structure ultrasound segmentation (CAMUS) and sub-EchoNet-Dynamic, which are two large-scale and public 2D echocardiography datasets. The experimental results show that PLANet performs better than traditional and deep learning-based segmentation methods on geometrical and clinical metrics. Moreover, PLANet can complete the segmentation of heart structures in 2D echocardiography in real time, indicating a potential to assist cardiologists accurately and efficiently.

Gastric cancer (GC) is the third leading cause of cancer-associated deaths globally. Accurate risk prediction of the overall survival (OS) for GC patients shows significant prognostic value, which helps identify and classify patients into different risk groups to benefit from personalized treatment. Many methods based on machine learning algorithms have been widely explored to predict the risk of OS. However, the accuracy of risk prediction has been limited and remains a challenge with existing methods. Few studies have proposed a framework and pay attention to the low-level and high-level features separately for the risk prediction of OS based on computed tomography images of GC patients. To achieve high accuracy, we propose a multi-focus fusion convolutional neural network. The network focuses on low-level and high-level features, where a subnet to focus on lower-level features and the other enhanced subnet with lateral connection to focus on higher-level semantic features. Three independent datasets of 640 GC patients are used to assess our method. Our proposed network is evaluated by metrics of the concordance index and hazard ratio. Our network outperforms state-of-the-art methods with the highest concordance index and hazard ratio in independent validation and test sets. Our results prove that our architecture can unify the separate low-level and high-level features into a single framework, and can be a powerful method for accurate risk prediction of OS.

Improving the quality of image-guided radiation therapy requires the tracking of respiratory motion in ultrasound sequences. However, the low signal-to-noise ratio and the artifacts in ultrasound images make it difficult to track targets accurately and robustly. In this study, we propose a novel deep learning model, called a Cascaded One-shot Deformable Convolutional Neural Network (COSD-CNN), to track landmarks in real time in long ultrasound sequences. Specifically, we design a cascaded Siamese network structure to improve the tracking performance of CNN-based methods. We propose a one-shot deformable convolution module to enhance the robustness of the COSD-CNN to appearance variation in a meta-learning manner. Moreover, we design a simple and efficient unsupervised strategy to facilitate the network's training with a limited number of medical images, in which many corner points are selected from raw ultrasound images to learn network features with high generalizability. The proposed COSD-CNN has been extensively evaluated on the public Challenge on Liver UltraSound Tracking (CLUST) 2D dataset and on our own ultrasound image dataset from the First Affiliated Hospital of Sun Yat-sen University (FSYSU). Experiment results show that the proposed model can track a target through an ultrasound sequence with high accuracy and robustness. Our method achieves new state-of-the-art performance on the CLUST 2D benchmark set, indicating its strong potential for application in clinical practice.

Fluorescence molecular tomography (FMT) is a highly sensitive and noninvasive imaging modality for three-dimensional visualization of fluorescence probe distribution in small animals. However, the simplified photon propagation model and ill-posed inverse problem limit the improvement of FMT reconstruction. In this work, we proposed a novel K-nearest neighbor based locally connected (KNN-LC) network to improve the performance of morphological reconstruction in FMT. It directly builds the inverse process of photon transmission by learning the mapping relation between the surface photon intensity and the distribution of fluorescent source. KNN-LC network cascades a fully connected (FC) sub-network with a locally connected (LC) sub-network, where the FC part provides a coarse reconstruction result and LC part fine-tunes the morphological quality of reconstructed result. To assess the performance of our proposed network, we implemented both numerical simulation and in vivo studies. Furthermore, split Bregman-resolved total variation (SBRTV) regularization method and inverse problem simulation (IPS) method were utilized as baselines in all comparisons. The results demonstrated that KNN-LC network achieved accurate reconstruction in both source localization and morphology recovery in a short time. This promoted the in vivo application of FMT for visualizing the distribution of biomarkers inside biological tissue.

Orientation Field (OF) is one of the most significant characters to distinguish fingerprint images from non-fingerprint images. An effective definition of fingerprint OF pattern will not only benefit fingerprint enhancement, but also contribute to latent fingerprint detection and segmentation. The existing fingerprint OF models either require pre-knowledge of singular points, or cannot be generalized to all kinds of fingerprint OFs. In this paper, we propose to define the fingerprint OF patterns based on low rank decomposition and sparse coding. Then we apply this proposed method to fingerprint OF recognition and detection. Experimental results prove the effectiveness of our method.

Overlapped fingerprints are commonly encountered in latent fingerprints lifted from crime scenes. Such overlapped fingerprints can hardly be processed by state-of-the-art fingerprint matchers. Several methods have been proposed to separate the overlapped fingerprints. However, these methods neither provide a robust separation results, nor could be generalized to most overlapped fingerprints. In this paper, we propose a novel overlapped fingerprint separation algorithm based on adaptive orientation model fitting. Different from existing methods, our algorithm estimates the initial orientation fields in a more accurate way and then separates the orientation fields for component fingerprints through an iterative correction process. Experimental results on latent overlapped fingerprints database demonstrate the advantage of our algorithm.

Overlapped fingerprints are commonly encountered in latent fingerprints lifted from crime scenes. Such overlapped fingerprints can hardly be processed by state-of-the-art fingerprint matchers. Several methods have been proposed to separate the overlapped fingerprints. However, these methods neither provide robust separation results, nor could be generalized for most overlapped fingerprints. In this paper, we propose a novel latent overlapped fingerprints separation algorithm based on adaptive orientation model fitting. Different from existing methods, our algorithm estimates the initial orientation fields in a more accurate way and then separates the orientation fields for component fingerprints through an iterative correction process. Global orientation field models are used to predict and correct the orientations in overlapped regions. Experimental results on the latent overlapped fingerprints database show that the proposed algorithm outperforms the state-of-the-art algorithm in terms of accuracy.

Accurate estimation of fingerprint orientation field is an essential module in fingerprint recognition. This paper proposes a novel technique for improving fingerprint orientation field estimation by fingerprint orientation model based on weighted 2D fourier expansion(W-FOMFE). The motivation for the proposed method can be found by: 1)the original FOMFE is sensitive to abrupt changes in orientation field; 2) blocks of different quality should have different impacts on FOMFE. Thus, we take into account the information of the Harris-corner strength (HCS) for orientation field estimation. In our method, we first calculate the fingerprint’ HCS; then use the HCS to remove abrupt changes in orientation field; finally, incorporate the normalized HCS as weighted value into original FOMFE. We test our method on FVC2004DB1. Experimental results show that our method (W-FOMFE) has better orientation field estimation than FOMFE.

The biometric template, which is stored in the form of raw data, has become the greatest potential threat to the security of biometric authentication system. As the compromise of the biometric data is permanent, the protection of biometric data is particularly important. Consequently, biometric template protection technologies have aroused research highlights recently. One of the most popular template protection methods is biometric cryptosystem method. In this paper, we design a codebook named distance projection for biometric coding to generate secured biometric template, and propose a novel fingerprint biometric cryptosystem scheme based on the codebook. Experimental results on FVC2002 DB2 show that the proposed scheme can obtain positive results on both security and authentication accuracy.

Because of the randomness of biometric features, traditional methods cannot perform well in the encryption of fingerprint template. The Fuzzy Vault construct is a biometric cryptosystem which can bind the fingerprint features with the secret key, and give reliable protection to both of them. The performance of Fuzzy Vault is highly affected by the alignment of fingerprints in the encryption field. In this paper, we propose a novel helper data based on the topo-structure around the core. The topo-structure has the advantage that it does not reveal any information about the template minutiae, and can reduce the alignment calculation amount. Experimental results show that the novel helper data has a good performance in alignment between vaults with high security.

This paper proposes a novel minutiae indexing method to speed up fingerprint matching, which narrows down the searching space of minutiae to reduce the expense of computation. An orderly sequence of features are extracted to describe each minutia and the indexing score is defined to select minutiae candidates from the query fingerprint for each minutia in the input fingerprint. The proposed method can be applied in both minutiae structure-based verification and fingerprint identification. Experiments are performed on a large-distorted fingerprint database (FVC2004 DB1) to approve the validity of the proposed method.

Current certificate-based information security technologies are facing a challenge of lacking the exact connection between cryptographic key and legitimate users and a problem of comprehensive management of certificates. In this paper we propose a novel networking identity authentication scheme based on fingerprint coding method and Identity-based encryption (IBE). Fingerprint code is a dual-factor authenticator which combines a token, e.g. USB key, with user's fingerprint feature by direct mixing of pseudo-random number generated by token and fingerprint feature. IBE allows for a sender to encrypt a message to a receiver without access to a public key certificate. This scheme can improve the security of user authentication and simplify comprehensive management of certificates, in the meantime, protect against fingerprint data fabrication and stolen. The security and efficiency of our proposed scheme meet the requirements of practical applications.

Most of secure email systems adopt PKI and IBE encryption schemes to meet security demands in communications via emails, however, both PKI and IBE encryption schemes have their own shortcomings and flaws and consequently bring security problems to email systems. This paper proposes a new secure email system based on a fingerprint authentication scheme which combines fingerprint authentication technology with IBE scheme. The system perfectly solves the existing problems encountered in email security protection implementations.

This paper introduces a robust fingerprint matching scheme based on the comprehensive minutia and the binary relation between minutiae. In the method, a fingerprint is represented as a graph, of which the comprehensive minutiae act as the vertex set and the local binary minutia relations provide the edge set. Then, the transformation-invariant and transformation-variant features are extracted from the binary relation. The transformation-invariant features are suitable to estimate the local matching probability, whereas the transformation-variant features are used to model the fingerprint rotation transformation with the adaptive Parzen window. Finally, the fingerprint matching is conducted with the variable bounded box method and iterative strategy. The experiments demonstrate that the proposed scheme is effective and robust in fingerprint alignment and matching.

This paper introduces a novel algorithm based on global comprehensive similarity with three steps. To describe the Euclidean space-based relative features among minutiae, we first build a minutia-simplex that contains a pair of minutiae as well as their associated textures, with its transformation-variant and invariant relative features employed for the comprehensive similarity measurement and parameter estimation, respectively. By the second step, we use the ridge-based nearest neighborhood among minutiae to represent the ridge-based relative features among minutiae. With these ridge-based relative features, minutiae are grouped according to their affinity with a ridge. The Euclidean space-based and ridge-based relative features among minutiae reinforce each other in the representation of a fingerprint. Finally, we model the relationship between transformation and the comprehensive similarity between two fingerprints in terms of histogram for initial parameter estimation. Through these steps, our experiment shows that the method mentioned above is both effective and suitable for limited memory AFIS owing to its less than 1k byte template size.

Coping with nonlinear distortions in fingerprint matching is a challenging task. This paper proposes a novel method, a fuzzy feature match (FFM) based on a local triangle feature set to match the deformed fingerprints. The fingerprint is represented by the fuzzy feature set: the local triangle feature set. The similarity between the fuzzy feature set is used to characterize the similarity between fingerprints. A fuzzy similarity measure for two triangles is introduced and extended to construct a similarity vector including the triangle-level similarities for all triangles in two fingerprints. Accordingly, a similarity vector pair is defined to illustrate the similarities between two fingerprints. The FFM method maps the similarity vector pair to a normalized value which quantifies the overall image to image similarity. The proposed algorithm has been evaluated with NIST 24 and FVC2004 fingerprint databases. Experimental results confirm that the proposed FFM based on the local triangle feature set is a reliable and effective algorithm for fingerprint matching with nonlinear distortions.

Coping with nonlinear distortions in fingerprint matching is a challenging task. This paper proposes a novel algorithm, normalized fuzzy similarity measure (NFSM), to deal with the nonlinear distortions. The proposed algorithm has two main steps. First, the template and input fingerprints were aligned. In this process, the local topological structure matching was introduced to improve the robustness of global alignment. Second, the method NFSM was introduced to compute the similarity between the template and input fingerprints. The proposed algorithm was evaluated on fingerprints databases of FVC2004. Experimental results confirm that NFSM is a reliable and effective algorithm for fingerprint matching with nonliner distortions. The algorithm gives considerably higher matching scores compared to conventional matching algorithms for the deformed fingerprints.

A biometric information processing toolkit (BITK) is designed and implemented to support the biometrics research, e.g. algorithm development and performance evaluation. BITK is an object-oriented C++ software development toolkit (SDK), and it provides a consistent, flexible and reusable framework to integrate algorithms, data structures, and visualization methods. In addition, an application platform based on BITK (BITKAPP) is developed. BITKAPP makes the best of BITK to serve the biometrics researchers with its friendly user interface and the plug-in architecture. The meaningful applications of the toolkit and platform confirm that they effectively support the biometrics research.

Abstract This paper introduces a fingerprint identification algorithm by clustering similarity with the view to overcome the dilemmas encountered in fingerprint identification. To decrease multi-spectrum noises in a fingerprint, we first use a dyadic scale space (DSS) method for image enhancement. The second step describes the relative features among minutiae by building a minutia-simplex which contains a pair of minutiae and their local associated ridge information, with its transformation-variant and invariant relative features applied for comprehensive similarity measurement and for parameter estimation respectively. The clustering method is employed to estimate the transformation space. Finally, multi-resolution technique is used to find an optimal transformation model for getting the maximal mutual information between the input and the template features. The experimental results including the performance evaluation by the 2nd International Verification Competition in 2002 (FVC2002), over the four fingerprint databases of FVC2002 indicate that our method is promising in an automatic fingerprint identification system (AFIS).