Showing 1–4 of 4 results for author: Wang, S Y

Common and Rare Fundus Diseases Identification Using Vision-Language Foundation Model with Knowledge of Over 400 Diseases

Authors: Meng Wang, Tian Lin, Aidi Lin, Kai Yu, Yuanyuan Peng, Lianyu Wang, Cheng Chen, Ke Zou, Huiyu Liang, Man Chen, Xue Yao, Meiqin Zhang, Binwei Huang, Chaoxin Zheng, Peixin Zhang, Wei Chen, Yilong Luo, Yifan Chen, Honghe Xia, Tingkun Shi, Qi Zhang, Jinming Guo, Xiaolin Chen, Jingcheng Wang, Yih Chung Tham , et al. (24 additional authors not shown)

Abstract: Previous foundation models for retinal images were pre-trained with limited disease categories and knowledge base. Here we introduce RetiZero, a vision-language foundation model that leverages knowledge from over 400 fundus diseases. To RetiZero's pre-training, we compiled 341,896 fundus images paired with text descriptions, sourced from public datasets, ophthalmic literature, and online resources… ▽ More Previous foundation models for retinal images were pre-trained with limited disease categories and knowledge base. Here we introduce RetiZero, a vision-language foundation model that leverages knowledge from over 400 fundus diseases. To RetiZero's pre-training, we compiled 341,896 fundus images paired with text descriptions, sourced from public datasets, ophthalmic literature, and online resources, encompassing a diverse range of diseases across multiple ethnicities and countries. RetiZero exhibits superior performance in several downstream tasks, including zero-shot disease recognition, image-to-image retrieval, and internal- and cross-domain disease identification. In zero-shot scenarios, RetiZero achieves Top5 accuracy scores of 0.8430 for 15 fundus diseases and 0.7561 for 52 fundus diseases. For image retrieval, it achieves Top5 scores of 0.9500 and 0.8860 for the same disease sets, respectively. Clinical evaluations show that RetiZero's Top3 zero-shot performance surpasses the average of 19 ophthalmologists from Singapore, China and the United States. Furthermore, RetiZero significantly enhances clinicians' accuracy in diagnosing fundus disease. These findings underscore the value of integrating the RetiZero foundation model into clinical settings, where a variety of fundus diseases are encountered. △ Less

Submitted 30 June, 2024; v1 submitted 13 June, 2024; originally announced June 2024.

Extraction of Text from Optic Nerve Optical Coherence Tomography Reports

Authors: Iyad Majid, Youchen Victor Zhang, Robert Chang, Sophia Y. Wang

Abstract: Purpose: The purpose of this study was to develop and evaluate rule-based algorithms to enhance the extraction of text data, including retinal nerve fiber layer (RNFL) values and other ganglion cell count (GCC) data, from Zeiss Cirrus optical coherence tomography (OCT) scan reports. Methods: DICOM files that contained encapsulated PDF reports with RNFL or Ganglion Cell in their document titles wer… ▽ More Purpose: The purpose of this study was to develop and evaluate rule-based algorithms to enhance the extraction of text data, including retinal nerve fiber layer (RNFL) values and other ganglion cell count (GCC) data, from Zeiss Cirrus optical coherence tomography (OCT) scan reports. Methods: DICOM files that contained encapsulated PDF reports with RNFL or Ganglion Cell in their document titles were identified from a clinical imaging repository at a single academic ophthalmic center. PDF reports were then converted into image files and processed using the PaddleOCR Python package for optical character recognition. Rule-based algorithms were designed and iteratively optimized for improved performance in extracting RNFL and GCC data. Evaluation of the algorithms was conducted through manual review of a set of RNFL and GCC reports. Results: The developed algorithms demonstrated high precision in extracting data from both RNFL and GCC scans. Precision was slightly better for the right eye in RNFL extraction (OD: 0.9803 vs. OS: 0.9046), and for the left eye in GCC extraction (OD: 0.9567 vs. OS: 0.9677). Some values presented more challenges in extraction, particularly clock hours 5 and 6 for RNFL thickness, and signal strength for GCC. Conclusions: A customized optical character recognition algorithm can identify numeric results from optical coherence scan reports with high precision. Automated processing of PDF reports can greatly reduce the time to extract OCT results on a large scale. △ Less

Submitted 21 August, 2023; originally announced August 2023.

Decoupling Control from Data for TCP Congestion Control

Authors: S. Y. Wang

Abstract: Many applications want to use TCP congestion control to regulate the transmission rate of a data packet stream. A natural way to achieve this goal is to transport the data packet stream on a TCP connection. However, because TCP implements both congestion and error control, transporting a data packet stream directly using a TCP connection forces the data packet stream to be subject to TCP's other… ▽ More Many applications want to use TCP congestion control to regulate the transmission rate of a data packet stream. A natural way to achieve this goal is to transport the data packet stream on a TCP connection. However, because TCP implements both congestion and error control, transporting a data packet stream directly using a TCP connection forces the data packet stream to be subject to TCP's other properties caused by TCP error control, which may be inappropriate for these applications. The TCP decoupling approach proposed in this thesis is a novel way of applying TCP congestion control to a data packet stream without actually transporting the data packet stream on a TCP connection. Instead, a TCP connection using the same network path as the data packet stream is set up separately and the transmission rate of the data packet stream is then associated with that of the TCP packets. Since the transmission rate of these TCP packets is under TCP congestion control, so is that of the data packet stream. Furthermore, since the data packet stream is not transported on a TCP connection, the regulated data packet stream is not subject to TCP error control. Because of this flexibility, the TCP decoupling approach opens up many new opportunities, solves old problems, and improves the performance of some existing applications. All of these advantages will be demonstrated in the thesis. This thesis presents the design, implementation, and analysis of the TCP decoupling approach, and its successful applications in TCP trunking, wireless communication, and multimedia streaming. △ Less

Submitted 21 October, 1999; originally announced October 1999.

Comments: Ph.D. Thesis, Harvard University, September 1999. This thesis's Chapter 3 about "TCP Trunking" will be published in IEEE ICNP'99 Proceedings. A condensed version of this thesis is currently submitted to a conference

ACM Class: C.2.2; C.2.5; C.2.6

TCP Trunking

Authors: H. T. Kung, S. Y. Wang

Abstract: A TCP trunk is an IP tunnel under TCP control, capable of carrying packets from any number of user flows. By exploiting properties of TCP, a TCP trunk provides elastic and reliable transmission over a network, and automatically shares the network fairly with other competing trunks. Moreover, by aggregating user flows into a single trunk flow, TCP trunking can significantly reduce the number of f… ▽ More A TCP trunk is an IP tunnel under TCP control, capable of carrying packets from any number of user flows. By exploiting properties of TCP, a TCP trunk provides elastic and reliable transmission over a network, and automatically shares the network fairly with other competing trunks. Moreover, by aggregating user flows into a single trunk flow, TCP trunking can significantly reduce the number of flows that the network needs to manage, thereby allowing use of simplified management to achieve improved perfor mance. For example, when dealing with only a small number of TCP trunk flows, a router with a simple FIFO buffer can experience low packet loss rates. A TCP trunk is a "soft" circuit in the sense that it requires no flow states to be maintained inside the network. Setting up a TCP trunk involves only configuring the two end nodes. This is in contrast with traditional methods of configuring circuits via signaling of network nodes. A simple packet-dropping mechanism based on packet accounting at the transmitter of a TCP trunk assures that, when the trunk reduces its bandwidth in response to network congestion, user TCP flows carried by the trunk will reduce their bandwidths by the same proportion. Simu lation results have demonstrated that TCP trunks can provide improved network performance to users, while achieving high network utilization. △ Less

Submitted 20 November, 1998; originally announced November 1998.

Comments: postscript file

ACM Class: C.2.1