Hongwei Han

Hongwei Han

Auto-generated profile, based on publicly available information

Hongwei Han

Tsinghua Shenzhen International Graduate School, Tsinghua University, Beijing, China

Research Focus & Metrics

Academic Career Length

6Years (2021-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
4 May 2022 in IEEE Transactions on Medical Imaging
Verified
Journal Article
Closed access
ZWZhanyu Wang
Zhanyu Wang
HHHongwei Han
Hongwei Han
LWLei Wang
Lei Wang
XLXiu Li
Xiu Li
LZLuping Zhou
Luping Zhou
Automated radiographic report generation is challenging in at least two aspects. First, medical images are very similar to each other and the visual differences of clinic importance are often fine-grained. Second, the disease-related words may be submerged by many similar sentences describing the common content of the images, causing the abnormal to be misinterpreted as the normal in the worst case. To tackle these challenges, this paper proposes a pure transformer-based framework to jointly enforce better visual-textual alignment, multi-label diagnostic classification, and word importance weighting, to facilitate report generation. To the best of our knowledge, this is the first pure transformer-based framework for medical report generation, which enjoys the capacity of transformer in learning long range dependencies for both image regions and sentence words. Specifically, for the first challenge, we design a novel mechanism to embed an auxiliary image-text matching objective into the transformer’s encoder-decoder structure, so that better correlated image and text features could be learned to help a report to discriminate similar images. For the second challenge, we integrate an additional multi-label classification task into our framework to guide the model in making correct diagnostic predictions. Also, a term-weighting scheme is proposed to reflect the importance of words for training so that our model would not miss key discriminative information. Our work achieves promising performance over the state-of-the-arts on two benchmark datasets, including the largest dataset MIMIC-CXR.

1 July 2021 in IOP Conference Series: Earth and Environmental Science
Verified
Conference Paper
Open access
PDF file
TZTianjiao Zhang
Tianjiao Zhang
YHYang Hu
Yang Hu
YFYuelin Feng
Yuelin Feng
YWYongshi Wang
Yongshi Wang
HHHongwei Han
Hongwei Han
XLXiaochen Li
Xiaochen Li
The buried depth of overpressured top seal layer in Niuzhuang sub-sag is mainly 2700–3600m, which is characterized by deep center and shallow edge in plane. Its lithology is a combination of dark grey, grey mudstone and sandy mudstone. There are a large number of carbonate cements and clay minerals in the argillaceous rocks that form the top sealing layer. The diagenesis of mudstone occurs in the rapid transformation period of illite/smectite mixed layer minerals, and corresponds to a large number of carbonate cements. The displacement pressure, thickness and diagenetic intensity of overpressured top seal are important factors affecting its physical sealing ability. There are five evolution stages of overpressured top seal, namely, at the initial stage of settlement, the strata are normally compacted, and the development of overpressured top seal is not obvious; During the first stage of accumulation, the low-amplitude hydrocarbon generation in the deep sag was supercharged, and the overpressured top seal was initially formed. Intermittent period of reservoir formation, overpressure top seal uplift erosion; During the rapid subsidence period, the strata are undercompacted to form overpressure, and the overpressure top seal is re-formed. During the second stage of storage, the thickness of overpressure top seal increased to the maximum.