The thickness of the choroid is considered to be an important indicator of clinical diagnosis. Therefore, accurate choroid segmentation in retinal OCT images is crucial for monitoring various ophthalmic diseases. However, this is still challenging due to the blurry boundaries and interference from other lesions. To address these issues, we propose a novel prior-guided and knowledge diffusive network (PGKD-Net) to fully utilize retinal structural information to highlight choroidal region features and boost segmentation performance. Specifically, it is composed of two parts: a Prior-mask Guided Network (PG-Net) for coarse segmentation and a Knowledge Diffusive Network (KD-Net) for fine segmentation. In addition, we design two novel feature enhancement modules, Multi-Scale Context Aggregation (MSCA) and Multi-Level Feature Fusion (MLFF). The MSCA module captures the long-distance dependencies between features from different receptive fields and improves the model's ability to learn global context. The MLFF module integrates the cascaded context knowledge learned from PG-Net to benefit fine-level segmentation. Comprehensive experiments are conducted to evaluate the performance of the proposed PGKD-Net. Experimental results show that our proposed method achieves superior segmentation accuracy over other state-of-the-art methods. Our code is made up publicly available at: https://github.com/yzh-hdu/choroid-segmentation.

Recently, object detection has experienced great progress. However, in the field of remote sensing, it is still under investigation for real-time detection of small objects because of the limited resolution and information. To detect a predefined object, such as aircraft, is important in some applications. This paper aims to improve the detection accuracy of the aircraft. Meanwhile, the speed of detection is another factor to be considered. At present, the model of You Only Look Once v5s (YOLOv5s) for object detection can achieve a good trade-off between accuracy and speed. So, YOLOv5s is adopted in the study. The model is trained with 600 remote-sensing images from UCAS-AOD dataset and tested with another 400 images from the same dataset. The experimental results show that YOLOv5 achieves 99.3% AP50in 74 seconds. It proves this model can effectively detect aircrafts in real time and achieve high accuracy.

This paper presents an instruction and Fabric Programmable Neuron Array (iFPNA) architecture, its 28nm CMOS chip prototype, and a compiler for the acceleration of a variety of deep learning neural networks (DNNs) including convolutional neural networks (CNNs), recurrent neural networks (RNNs), and fully connected (FC) networks on chip. The iFPNA architecture combines instruction-level programmability as in an Instruction Set Architecture (ISA) with logic-level reconfigurability as in a Field-Programmable Gate Array (FPGA) in a sliced structure for scalability. Four data flow models, namely weight stationary, input stationary, row stationary and tunnel stationary, are described as the abstraction of various DNN data and computational dependence. The iFPNA compiler partitions a large-size DNN to smaller networks, each being mapped to, optimized and code generated for, the underlying iFPNA processor using one or a mixture of the four data-flow models. Experimental results have shown that state-of-art large-size CNNs, RNNs, and FC networks can be mapped to the iFPNA processor achieving the near ASIC performance.

Redirected walking techniques have the potential to provide natural locomotion while users experience large virtual environments. However, when using redirected walking in small physical workspaces, disruptive overt resets are often required. We describe the design of an educational virtual reality experience in which users physically walk through virtual tunnels representative of the World War I battle of Vauquois. Walking in only a 15- by 5-foot tracked space, users are redirected through subtle, narrative-driven resets to walk through a tunnel nearly 50 feet in length. This work contributes approaches and lessons that can be used to provide a seamless and natural virtual reality walking experience in highly constrained physical spaces.

In this paper, we present a fingerprint-based particle filtering (PF) approach for the mobility tracking in cellular networks. With the popularity of location-based services (LBSs) in a recent decade, the mobility tracking has now become one of the indispensable techniques to fulfill the pervasive and location-aware computing. However, the tracking results from the conventional K nearest neighbors (KNN) and Kalman filtering (KF) by the cellular data appear to be coarse due to the reflection, refraction and diffraction of received signal code power (RSCP). Therefore, we propose the KNN/PF as an effective way to improve the stability and accuracy of mobility tracking in cellular networks. Furthermore, the experiments conducted in Pudong New District, Shanghai, China, show that the KNN/PF approach can achieve better cumulative density function (CDF) of tracking errors compared with the KNN and the mixed KNN and KF (KNN/KF) approaches.