A modified U-Net convolutional neural network for segmenting periprostatic adipose tissue based on contour feature learning.

Objective: This study trains a U-shaped fully convolutional neural network (U-Net) model based on peripheral contour measures to achieve rapid, accurate, automated identification and segmentation of periprostatic adipose tissue (PPAT). Methods: Currently, no studies are using deep learning methods to discriminate and segment periprostatic adipose tissue. This paper proposes a novel and modified, U-shaped convolutional neural network contour control points on a small number of datasets of MRI T2W images of PPAT combined with its gradient images as a feature learning method to reduce feature ambiguity caused by the differences in PPAT contours of different patients. This paper adopts a supervised learning method on the labeled dataset, combining the probability and spatial distribution of control points, and proposes a weighted loss function to optimize the neural network's convergence speed and detection performance. Based on high-precision detection of control points, this paper uses a convex curve fitting to obtain the final PPAT contour. The imaging segmentation results were compared with those of a fully convolutional network (FCN), U-Net, and semantic segmentation convolutional network (SegNet) on three evaluation metrics: Dice similarity coefficient (DSC), Hausdorff distance (HD), and intersection over union ratio (IoU). Results: Cropped images with a 270 × 270-pixel matrix had DSC, HD, and IoU values of 70.1%, 27 mm, and 56.1%, respectively; downscaled images with a 256 × 256-pixel matrix had 68.7%, 26.7 mm, and 54.1%. A U-Net network based on peripheral contour characteristics predicted the complete periprostatic adipose tissue contours on T2W images at different levels. FCN, U-Net, and SegNet could not completely predict them. Conclusion: This U-Net convolutional neural network based on peripheral contour features can identify and segment periprostatic adipose tissue quite well. Cropped images with a 270 × 270-pixel matrix are more appropriate for use with the U-Net convolutional neural network based on contour features; reducing the resolution of the original image will lower the accuracy of the U-Net convolutional neural network. FCN and SegNet are not appropriate for identifying PPAT on T2 sequence MR images. Our method can automatically segment PPAT rapidly and accurately, laying a foundation for PPAT image analysis.

Environment-Sensitive Fluorescent Probe Enables Assessment of Anaplastic Lymphoma Kinase Activity in Nonsmall Cell Lung Cancer.

Anaplastic lymphoma kinase (ALK) rearrangements have been identified as key oncogenic drivers of a subset of nonsmall cell lung cancer (NSCLC). The final chimeric protein of the fusion gene can be constitutively activated, which accounts for the growth and proliferation of ALK-rearranged tumors and thus strongly associates with cancer invasion and metastasis. Diagnostic tools enabling the visualization of ALK activity in a structure-function-based approach are highly desirable to determine ALK status and guide ALK tyrosine kinase inhibitor (ALK-TKI) treatment making. Here, we describe the design, synthesis, and application of a new environment-sensitive fluorescent probe HX16 by introducing an environment-sensitive fluorophore 4-sulfonamidebenzoxadiazole to visualize ALK activity in living cancer cells and tumor tissue slices (mouse model and human biopsy sample). HX16 is a multifunctional chemical tool based on the pharmacophore of ALK-TKI (ceritinib) and can specifically target the kinase domain of ALK with a high sensitivity. Using flow cytometry and confocal microscopy, HX16 enables visualization of ALK activity in various cancer cells with distinct ALK fusion genes, as well as xenograft mouse models. Importantly, HX16 was also applied to visualize ALK activity in a tumor biopsy from a NSCLC patient with ALK-echinoderm microtubule-associated protein-like-4 fusion gene for prediction of ALK-TKI sensitivity. These results demonstrate that strategically designed ALK-TKI-based probe allows the assessment of ALK activity in tumor tissues and hold promise as a useful diagnostic tool in predicting ALK-TKI therapy response.

Therapeutic potential of targeting polo-like kinase 4.

Polo-like kinase 4 (PLK4), a highly conserved serine/threonine kinase, masterfully regulates centriole duplication in a spatiotemporal manner to ensure the fidelity of centrosome duplication and proper mitosis. Abnormal expression of PLK4 contributes to genomic instability and associates with a poor prognosis in cancer. Inhibition of PLK4 is demonstrated to exhibit significant efficacy against various types of human cancers, further highlighting its potential as a promising therapeutic target for cancer treatment. As such, numerous small-molecule inhibitors with distinct chemical scaffolds targeting PLK4 have been extensively investigated for the treatment of different human cancers, with several undergoing clinical evaluation (e.g., CFI-400945). Here, we review the structure, distribution, and biological functions of PLK4, encapsulate its intricate regulatory mechanisms of expression, and highlighting its multifaceted roles in cancer development and metastasis. Moreover, the recent advancements of PLK4 inhibitors in patent or literature are summarized, and their therapeutic potential as monotherapies or combination therapies with other anticancer agents are also discussed.

An optical multiple-image authentication based on computational ghost imaging and total-variation minimization.

An optical multiple-image authentication is suggested using computational ghost imaging and total-variation minimization. Differing from encrypting multiple images into a noise-like ciphertext directly, as described in most conventional authentication methods, the related encoded information is embedded into a cover image to avoid the attention of eavesdroppers. First, multiple images are encoded to form real-valued sequences composed of corresponding bucket values obtained by the aid of computational ghost imaging, and four sub-images are obtained by decomposing the cover image using wavelet transform. Second, measured sequences are embedded into one of the sub-images, and embedding positions are randomly selected using corresponding binary masks. To enhance the security level, a chaotic sequence is produced using logistic map and used to scramble measured intensities. Most importantly, original images with high quality can be directly recovered using total-variation minimization. The validity and robustness of the proposed approach are verified with optical experiments.

Optical multiple-image authentication based on computational ghost imaging and hybrid non-convex second-order total variation.

An optical security method for multiple-image authentication is proposed based on computational ghost imaging and hybrid non-convex second-order total variation. Firstly, each original image to be authenticated is encoded to the sparse information using computational ghost imaging, where illumination patterns are generated based on Hadamard matrix. In the same time, the cover image is divided into four sub-images with wavelet transform. Secondly, one of sub-images with low-frequency coefficients is decomposed using singular value decomposition (SVD), and all sparse data are embedded into the diagonal matrix with the help of binary masks. To enhance the security, the generalized Arnold transform is used to scramble the modified diagonal matrix. After using SVD again, the marked cover image carrying the information of multiple original images is obtained using the inverse wavelet transform. In the authentication process, the quality of each reconstructed image can be greatly improved based on hybrid non-convex second-order total variation. Even at a very low sampling ratio (i.e., 6%), the existence of original images can be efficiently verified using the nonlinear correlation maps. To our knowledge, it is first to embed sparse data into the high-frequency sub-image using two cascaded SVDs, which can guarantee high robustness against the Gaussian filter and sharpen filter. The optical experiments demonstrate the feasibility of the proposed mechanism, which can provide an effective alternative for the multiple-image authentication.

Developing non-invasive 3D quantificational imaging for intelligent coconut analysis system with X-ray.

BACKGROUND: As one of the largest drupes in the world, the coconut has a special multilayered structure and a seed development process that is not yet fully understood. On the one hand, the special structure of the coconut pericarp prevents the development of external damage to the coconut fruit, and on the other hand, the thickness of the coconut shell makes it difficult to observe the development of bacteria inside it. In addition, coconut takes about 1 year to progress from pollination to maturity. During the long development process, coconut development is vulnerable to natural disasters, cold waves, typhoons, etc. Therefore, nondestructive observation of the internal development process remains a highly important and challenging task. In this study, We proposed an intelligent system for building a three-dimensional (3D) quantitative imaging model of coconut fruit using Computed Tomography (CT) images. Cross-sectional images of coconut fruit were obtained by spiral CT scanning. Then a point cloud model was built by extracting 3D coordinate data and RGB values. The point cloud model was denoised using the cluster denoising method. Finally, a 3D quantitative model of a coconut fruit was established. RESULTS: The innovations of this work are as follows. 1) Using CT scans, we obtained a total of 37,950 non-destructive internal growth change maps of various types of coconuts to establish a coconut data set called "CCID", which provides powerful graphical data support for coconut research. 2) Based on this data set, we built a coconut intelligence system. By inputting a batch of coconut images into a 3D point cloud map, the internal structure information can be ascertained, the entire contour can be drawn and rendered according to need, and the long diameter, short diameter and volume of the required structure can be obtained. We maintained quantitative observation on a batch of local Hainan coconuts for more than 3 months. With 40 coconuts as test cases, the high accuracy of the model generated by the system is proven. The system has a good application value and broad popularization prospects in the cultivation and optimization of coconut fruit. CONCLUSION: The evaluation results show that the 3D quantitative imaging model has high accuracy in capturing the internal development process of coconut fruits. The system can effectively assist growers in internal developmental observations and in structural data acquisition from coconut, thus providing decision-making support for improving the cultivation conditions of coconuts.

A Light Field FDL-HCGH Feature in Scale-Disparity Space.

Many computer vision applications rely on feature detection and description, hence the need for computationally efficient and robust 4D light field (LF) feature detectors and descriptors. In this paper, we propose a novel light field feature descriptor based on the Fourier disparity layer representation, for light field imaging applications. After the Harris feature detection in a scale-disparity space, the proposed feature descriptor is then extracted using a circular neighborhood rather than a square neighborhood. It is shown to yield more accurate feature matching, compared with the LiFF LF feature, with a lower computational complexity. In order to evaluate the feature matching performance with the proposed descriptor, we generated a synthetic stereo LF dataset with ground truth matching points. Experimental results with synthetic and real-world dataset show that our solution outperforms existing methods in terms of both feature detection robustness and feature matching accuracy.

Expanded Application of a Photoaffinity Probe to Study Epidermal Growth Factor Receptor Tyrosine Kinase with Functional Activity.

The abnormal activation of the epidermal growth factor receptor (EGFR) is strongly associated with cancer invasion and metastasis. Tools and methods are required to study and visualize EGFR activation under (patho)physiological conditions. Here, we report the development of a two-step photoaffinity probe (HX101) by incorporation of a diazirine as a photoreactive group and an alkyne as a ligation handle to quantitively study EGFR kinase activity in native cellular contexts and human tissue slices. HX101 is a multifunctional probe based on the pharmacophore of the EGFR tyrosine kinase inhibitor (EGFR-TKI) and can covalently target the EGFR upon photoactivation. The incorporated alkyne serves as a versatile ligation handle and enables HX101 to introduce distinct reporter groups (e.g., fluorophore and biotin) via click chemistry. With variable reporter tags, HX101 enables visualization and target engagement studies of the active EGFR in a panel of cancer cells using flow cytometry, confocal microscopy, and mass spectrometry. Furthermore, as a proof of concept study, we applied HX101 in stochastic optical reconstruction microscopy super-resolution imaging to study EGFR activation in live cells. Importantly, HX101 was also applied to visualize EGFR mutant activity in tumor tissues from lung cancer patients for prediction of EGFR-TKI sensitivity. Altogether, our results demonstrate the wide application of a selective photoaffinity probe in multi-modal assessment/visualization of EGFR activity in both live cells and tissue slices. We anticipate that these diverse applications can facilitate the translation of a strategically functionalized probe into medical use.

Classification for Human Balance Capacity Based on Visual Stimulation under a Virtual Reality Environment.

The normal and disordered people balance ability classification is a key premise for rehabilitation training. This paper proposes a multi-barycentric area model (MBAM), which can be applied for accurate video analysis based classification. First, we have invited fifty-three subjects to wear an HTC (High Tech Computer Corporation) VIVE (Very Immersive Virtual Experience) helmet and to walk ten meters while seeing a virtual environment. The subjects' motion behaviors are collected as our balance ability classification dataset. Secondly, we use background differential algorithm and bilateral filtering as the preprocessing to alleviate the video noise and motion blur. Inspired by the balance principle of a tumbler, we introduce a MBAM model to describe the body balancing condition by computing the gravity center of a triangle area, which is surrounded by the upper, middle and lower parts of the human body. Finally, we can obtain the projection coordinates according to the center of gravity of the triangle, and get the roadmap of the subjects by connecting those projection coordinates. In the experiments, we adopt four kinds of metrics (the MBAM, the area variance, the roadmap and the walking speed) innumerical analysis to verify the effect of the proposed method. Experimental results show that the proposed method can obtain a more accurate classification for human balance ability. The proposed research may provide potential theoretical support for the clinical diagnosis and treatment for balance dysfunction patients.

Aliasing Detection and Reduction Scheme on Angularly Undersampled Light Fields.

When using plenoptic camera for digital refocusing, angular undersampling can cause severe (angular) aliasing artifacts. Previous approaches have focused on avoiding aliasing by pre-processing the acquired light field via prefiltering, demosaicing, reparameterization, and so on. In this paper, we present a different solution that first detects and then removes angular aliasing at the light field refocusing stage. Different from previous frequency domain aliasing analysis, we carry out a spatial domain analysis to reveal whether the angular aliasing would occur and uncover where in the image it would occur. The spatial analysis also facilitates easy separation of the aliasing versus non-aliasing regions and angular aliasing removal. Experiments on both synthetic scene and real light field data sets (camera array and Lytro camera) demonstrate that our approach has a number of advantages over the classical prefiltering and depth-dependent light field rendering techniques.