Retrosynthesis and forward synthesis prediction are fundamental challenges in organic synthesis, computer-aided synthesis planning (CASP), and computer-aided drug design (CADD). The objective is to predict plausible reactants for a given target product and its corresponding inverse task. With the rapid development of deep learning, numerous approaches have been proposed to solve this problem from various perspectives. The methods based on molecular graphs benefit from their rich features embedded inside but face difficulties in applying existing sequence-based data augmentations due to the permutation invariance of graph structures. In this work, we propose SeqAGraph, a template-free approach that annotates input graphs with its root atom index to ensure compatibility with sequence-based data augmentation. The matrix product for global attention in graph encoders is implemented by indexing, elementwise product, and aggregation to fuse global attention with local message passing without graph padding. Experiments demonstrate that SeqAGraph fully benefits from molecular graphs and sequence-based data augmentation and achieves state-of-the-art accuracy in template-free approaches.

Stable and reliable operation of high-speed railway requires continuous and reliable power supply of traction substation, and isolating switch is one of the most important electrical devices in high-speed railway traction substation. In this paper, to address the issue of isolating switch accurate localization and state recognition simultaneously, we present an automatic isolating switch segmentation and state recognition framework called ISSSR-Net using multitask learning that consists of two stages. First, an isolating switch segmentation network called ISS-Net is proposed for isolating switch pixel-level segmentation precisely, of which a new structure containing a strip pooling module, a channel attention and three pyramid pooling modules is designed to greatly improve the segmentation and recognition performance, even in complex conditions such as rain, snow and fog. Second, to improve state recognition accuracy, the segmentation map yielded from the ISS-Net and the feature map from the shared backbone are together fed into isolating switch recognition network called ISR-Net to recognize its three states. In addition, a global context block is integrated into ISR-Net to further improve state recognition accuracy. Extensive experimental results on a self-collected dataset from Heishan traction substation corroborate that this paper provides an effective and robust method to achieve isolating switch segmentation and state recognition simultaneously. The MIoUand${F}~1$-score of segmentation reach 0.93 and 0.94 respectively, which is better than U-Net and its variants, and other SOTA. The${F}~1$-score of recognition reach 1.00, which is higher than HOG+SVM and other deep learning methods that have been experimented.

Spine surgeries are high risk operations which require the surgeons to have ample experiences. For young surgeons, effective and extensive training is critical. This paper presents a real time haptic spine surgical simulator that will be used to train residents, fellows and spine surgeons in a hospital training program. It provides a realistic environment for the trainees to practice spine surgeries and has the advantages of being interactive, low-cost, representative, and repeatable over conventional training approaches. Haptic Phantom offers the users force feedback, differentiating our system from other screen-based training systems. Computational efficiency was achieved by developing advanced graphical rendering methods. The volumetric data was classified into surface voxel cloud and inner voxel cloud by the adjacency graph which stored the relationship among voxels. To speed up the collision detection and real time rendering between the virtual surgical tools and the lumbar model, Octree-based algorithms and GPU technique were applied. To enhance the physical realism, three dimensional lumbar vertebrae models were reconstructed from CT images and associated with non-homogeneous bone density such that the rendered model best represents the spine anatomy and mechanics. We demonstrate system performance by conducting pedicle screw insertion.

Accurate measurement of lower limb alignment is vital for orthopedic surgery planning. Based on the 3D lower limb bone models reconstructed from CT images, a 3D feature construction system is developed to find the lower limb alignment. This paper presents new methods to automatically identify the anatomic axes, mechanical axes, joint center points, and reference planes of the lower extremities. As far as we know, there is no other system that can construct all these features automatically. With the above identified features, our system can calculate the related angles of the lower limb alignment accurately for orthopedic knee surgery planning. We applied our system on two groups of specimen (with 6 pairs of lower limbs), and the resulting related angles are identical to the actual angles of the lower limb alignment.

Previous ITiCSE papers have reported on Network EducationWare (NEW), an open source software system that supports synchronous and asynchronous distance education easily and inexpensively via the Internet. This paper reports on an innovative capability recently added to NEW that enables simultaneous execution of applications on all participating computers in both Windows and Linux systems, and explains how we have used this capability to teach a course in Computer Graphics online. Teaching graphics in this way previously was impossible because transmitting the visual output of graphics programs in real time with good quality is too demanding for normal Internet connections. This problem was solved by using the NEW application launcher to invoke Java classes, provided by the instructor, on all student computers simultaneously. The capability works with modest Internet capacity and also is captured in recorded sessions for asynchronous use. We explain how the NEW capability works and describe its use in online teaching of Computer Graphics.

Abstract We present a new post processing method of simulating depth of field based on accurate calculations of circles of confusion. Compared to previous work, our method derives actual scene depth information directly from the existing depth buffer, requires no specialized rendering passes, and allows easy integration into existing rendering applications. Our implementation uses an adaptive, two‐pass filter, producing a high quality depth of field effect that can be executed entirely on the GPU, taking advantage of the parallelism of modern graphics cards and permitting real time performance when applied to large numbers of pixels.

In this paper, we present an X-ray radiography system which can provide intra-operative stereo imaging of the patient's operational area during a spinal surgery. We use two horizontally separated X-ray tubes and a flat-panel X-ray detector mounted on two ends of a C-arm microcontroller synchronizes the image acquisition, and the left and right images are grabbed into the system alternately. Surgeons can have a 3D perception of the spine's X-ray image in real-time. We further developed a whole image processing language called PPL. PPL takes images as operands instead of dealing with complex pixel operation. To speed up the rendering speed in a network environment, we proposed a level of detail (LOD) with reverse prediction method that can fast transmit images over Internet in real-time.

This paper summarizes our NSF funded project, a PC-based multi-user learning environment: Multi-User Virtual Environment Experiential Simulator (MUVEES). The goal of this project is to create and evaluate graphical multi-user virtual environments that use digitized museum resources to enhance middle school students' motivation and learning about science. Here, we discuss the design, implementation, and applications of MUVEES. We present its structure, efficient approaches that achieve more realistic avatar behaviors, and pedagogical strategies that foster strong learning outcomes across a wide range of individual student characteristics. Our preliminary results indicate that MUVEES is a powerful vehicle for collaboration and learning. We believe that our system and implementation methods will help improve future multi-user virtual environments.

This paper describes the design and development of a VR system that can assist orthopedic diagnosis and surgery. A realistic 3D knee surface model, integrated with motion analysis, is used to visualize the geometrical and biomechanical characteristics of human knee joint. The system can be used to perform pre-operative surgery simulation and evaluation in a VR environment.

This paper presents an approximate multiple segment line scan‐conversion method — the Slope Table Method. The statistics show that the new method can increase the percentage of multiple segment lines (i.e., lines with more than one segment) in an N×N raster area from about 39% to more than 99%. In software implementation for scan‐conversion and antialiasing of randomly generated lines, this method is on average more than 6 times faster than Gupta's antialiasing line algorithm. Compared with other line scan‐conversion methods, the method may choose pixels which are not the closest to the line (i.e., error pixels). Here the paper demonstrates that the visual effect is acceptable in most applications with the error pixels.

We present a new method for scan‐converting a straight line. A scan‐converted straight line may contain many pixel segments of identical shapes. Therefore, instead of scan‐converting the whole line step by step, we can scan‐convert multiple segments of a line through copying and replicating. With a few modifications to the existing algorithms, we can significantly speed up the scan‐conversion process. The algorithm can be implemented in the hardware, further speeding up the process. The hardware design for the new algorithm is provided. We prove that we can speed up all existing line scan‐conversion algorithms on the average 3 times.