Automatic diabetic retinopathy (DR) lesions segmentation makes great sense of assisting ophthalmologists in diagnosis. Although many researches have been conducted on this task, most prior works paid too much attention to the designs of networks instead of considering the pathological association for lesions. Through investigating the pathogenic causes of DR lesions in advance, we found that certain lesions are closed to specific vessels and present relative patterns to each other. Motivated by the observation, we propose a relation transformer block (RTB) to incorporate attention mechanisms at two main levels: a self-attention transformer exploits global dependencies among lesion features, while a cross-attention transformer allows interactions between lesion and vessel features by integrating valuable vascular information to alleviate ambiguity in lesion detection caused by complex fundus structures. In addition, to capture the small lesion patterns first, we propose a global transformer block (GTB) which preserves detailed information in deep network. By integrating the above blocks of dual-branches, our network segments the four kinds of lesions simultaneously. Comprehensive experiments on IDRiD and DDR datasets well demonstrate the superiority of our approach, which achieves competitive performance compared to state-of-the-arts.

The perception of drones, also known as Unmanned Aerial Vehicles (UAVs), particularly in infrared videos, is crucial for effective anti-UAV tasks. However, existing datasets for UAV tracking have limitations in terms of target size and attribute distribution characteristics, which do not fully represent complex realistic scenes. To address this issue, we introduce a generalized infrared UAV tracking benchmark called Anti-UAV410 . The benchmark comprises a total of 410 videos with over 438 K manually annotated bounding boxes. To tackle the challenges of UAV tracking in complex environments, we propose a novel method called Siamese drone tracker ( SiamDT ). SiamDT incorporates a dual-semantic feature extraction mechanism that explicitly models targets in dynamic background clutter, enabling effective tracking of small UAVs. The SiamDT method consists of three key steps: Dual-Semantic RPN Proposals (DS-RPN) , Versatile R-CNN (VR-CNN) , and Background Distractors Suppression . These steps are responsible for generating candidate proposals, refining prediction scores based on dual-semantic features, and enhancing the discriminative capacity of the trackers against dynamic background clutter, respectively. Extensive experiments conducted on the Anti-UAV410 dataset and three other large-scale benchmarks demonstrate the superior performance of the proposed SiamDT method compared to recent state-of-the-art trackers.

Modeling layout is an important first step for graphic design. Recently, methods for generating graphic layouts have progressed, particularly with Generative Adversarial Networks (GANs). However, the problem of specifying the locations and sizes of design elements usually involves constraints with respect to element attributes, such as area, aspect ratio and reading-order. Automating attribute conditional graphic layouts remains a complex and unsolved problem. In this paper, we introduce Attribute-conditioned Layout GAN to incorporate the attributes of design elements for graphic layout generation by forcing both the generator and the discriminator to meet attribute conditions. Due to the complexity of graphic designs, we further propose an element dropout method to make the discriminator look at partial lists of elements and learn their local patterns. In addition, we introduce various loss designs following different design principles for layout optimization. We demonstrate that the proposed method can synthesize graphic layouts conditioned on different element attributes. It can also adjust well-designed layouts to new sizes while retaining elements' original reading-orders. The effectiveness of our method is validated through a user study.

With the continuous progress of development, deep learning has made good progress in the analysis and recognition of images, which has also triggered some researchers to explore the area of combining deep learning with hyperspectral medical images and achieve some progress. This paper introduces the principles and techniques of hyperspectral imaging systems, summarizes the common medical hyperspectral imaging systems, and summarizes the progress of some emerging spectral imaging systems through analyzing the literature. In particular, this article introduces the more frequently used medical hyperspectral images and the pre-processing techniques of the spectra, and in other sections, it discusses the main developments of medical hyperspectral combined with deep learning for disease diagnosis. On the basis of the previous review, tne limited factors in the study on the application of deep learning to hyperspectral medical images are outlined, promising research directions are summarized, and the future research prospects are provided for subsequent scholars.

Detecting small objects in drone imagery is challenging due to low resolution and background blending, leading to limited feature information. Multi-scale feature fusion can enhance detection by capturing information at different scales, but traditional strategies fall short. Simple concatenation or addition operations do not fully utilize multi-scale fusion advantages, resulting in insufficient correlation between features. This inadequacy hinders the detection of small objects, especially in complex backgrounds and densely populated areas. To address this issue and efficiently utilize the limited computational resources, we propose a lightweight fusion strategy based on enhanced inter-layer feature correlation (EFC) to replace the traditional feature fusion strategy in FPN. The semantic expressions of different layers in the feature pyramid are inconsistent. In EFC, the Grouped Feature Focus Unit (GFF) enhances the feature correlation of each layer by focusing on the contextual information of different features. The Multi-Level Feature Reconstruction Module (MFR) effectively reconstructs and transforms the strength and weakness information of each layer in the pyramid to reduce redundant feature fusion and retain more information about small targets in deep networks. It is noteworthy that the proposed method is plug-and-play and can be widely applied to various base networks. Extensive experiments and comprehensive evaluations on VisDrone, UAVDT and COCO demonstrate the effectiveness. Using GFL as the baseline on the VisDrone dataset with a large number of small targets, the proposed method improves the detection mAP by 1.7%, surpassing many lightweight state-of-the-art methods and significantly reducing the Params and GFLOPs at the neck end.

Fourier ptychographic microscopy (FPM) is a newly developed microscopic technique for large field of view, high-resolution and quantitative phase imaging by combining the techniques from ptychographic imaging, aperture synthesizing and phase retrieval. In FPM, an LED array is utilized to illuminate the specimen from different angles and the corresponding intensity images are synthesized to reconstruct a high-resolution complex field. As a flexible and low-cost approach to achieve high-resolution, wide-field and quantitative phase imaging, FPM is of enormous potential in biomedical applications such as hematology and pathology. Conventionally, the FPM reconstruction problem is solved by using a phase retrieval method, termed Alternate Projection. By iteratively updating the Fourier spectrum with low-resolution-intensity images, the result converges to a high-resolution complex field. Here we propose a new FPM reconstruction framework with deep learning methods and design a multiscale, deep residual neural network for FPM reconstruction. We employ the widely used open-source deep learning library PyTorch to train and test our model and carefully choose the hyperparameters of our model. To train and analyze our model, we build a large-scale simulation dataset with an FPM imaging model and an actual dataset captured with an FPM system. The simulation dataset and actual dataset are separated as training datasets and test datasets, respectively. Our model is trained with the simulation training dataset and fine tuned with the fine-tune dataset, which contains actual training data. Both our model and the conventional method are tested on the simulation test dataset and the actual test dataset to evaluate the performances. We also show the reconstruction result of another neural network-based method for comparison. The experiments demonstrate that our model achieves better reconstruction results and consumes much less time than conventional methods. The results also point out that our model is more robust under system aberrations such as noise and blurring (fewer intensity images) compared with conventional methods.