Ka Zhang

Prof. Ka Zhang

Nanjing Normal University, Nanjing, Jiangsu, China
orcid

Research Focus & Metrics

Academic Career Length

15Years (2012-Present)

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

Representative Publications

View All Publications
15 January 2021 in Journal of Cloud Computing
Verified
Journal Article
Open access
PDF file
JWJunshu Wang
Junshu Wang
GZGuoming Zhang
Guoming Zhang
WWWei Wang
Wei Wang
Ka ZhangKa Zhang
Ka Zhang
YSYehua Sheng
Yehua Sheng
With the rapid development of hospital informatization and Internet medical service in recent years, most hospitals have launched online hospital appointment registration systems to remove patient queues and improve the efficiency of medical services. However, most of the patients lack professional medical knowledge and have no idea of how to choose department when registering. To instruct the patients to seek medical care and register effectively, we proposed CIDRS, an intelligent self-diagnosis and department recommendation framework based on Chinese medical Bidirectional Encoder Representations from Transformers (BERT) in the cloud computing environment. We also established a Chinese BERT model (CHMBERT) trained on a large-scale Chinese medical text corpus. This model was used to optimize self-diagnosis and department recommendation tasks. To solve the limited computing power of terminals, we deployed the proposed framework in a cloud computing environment based on container and micro-service technologies. Real-world medical datasets from hospitals were used in the experiments, and results showed that the proposed model was superior to the traditional deep learning models and other pre-trained language models in terms of performance.

28 October 2022 in IEEE Transactions on Geoscience and Remote Sensing
Verified
Journal Article
Closed access
HCHui Chen
Hui Chen
LCLiang Cheng
Liang Cheng
QZQizhi Zhuang
Qizhi Zhuang
Ka ZhangKa Zhang
Ka Zhang
NLNing Li
Ning Li
ZDZhixin Duan
Zhixin Duan
The use of fully supervised deep learning methods to extract buildings from remote sensing images has shown excellent performance, which requires large amounts of training data with laborious per-pixel labeling. Compared with pixelated intensive labeling, it is much easier to label data using scribbles, which only takes few seconds for one image. In this paper, a one-stage structure-aware weakly supervised network (SAWSN) for building extraction is proposed, and it learns from easily accessible scribbles rather than from densely annotated ground truth. Firstly, to solve the problem that direct training with scribble labels will lead to poor building structures, an auxiliary edge detection task is introduced to localize building edges explicitly. Secondly, a structure aware scribble extension module (SASEM) is designed to recover building structures from scribbles through effective utilization of edge features. Finally, an edge-structure-aware loss is proposed to limit the scope of the restored structure. We perform extensive experiments on three newly labeled benchmark building extraction datasets (WHU, ISPRS Potsdam, and Vaihingen). Experimental results show that our method achieved 91.72%, 92.83%, and 92.22% of F1 using the ISPRS Vaihingen, Potsdam, and WHU datasets, respectively, and outperformed the state-of-the-art scribble-based weakly supervised methods by 3.27% of IoU .

22 January 2020 in IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing
Verified
Journal Article
Open access
YTYijie Tang
Yijie Tang
QWQunming Wang
Qunming Wang
Ka ZhangKa Zhang
Ka Zhang
PAPeter Atkinson
Peter Atkinson
Abstract—It is challenging to acquire satellite sensor data with both fine spatial and fine temporal resolution, especially for monitoring at global scales. Amongst the widely used global monitoring satellite sensors, Landsat data have a coarse temporal resolution, but fine spatial resolution, while MODIS data have fine temporal resolution, but coarse spatial resolution. One solution to this problem is to blend the two types of data using spatio-temporal fusion, creating images with both fine temporal and fine spatial resolution. However, reliable geometric registration of images acquired by different sensors is a prerequisite of spatio-temporal fusion. Due to the potentially large differences between the spatial resolutions of the images to be fused, the geometric registration process always contains some degree of uncertainty. This paper analyzes quantitatively the influence of geometric registration error on spatio-temporal fusion. The relationship between registration error and the accuracy of fusion was investigated under the influence of different temporal distances between images, different spatial patterns within the images and using different methods (i.e., spatial and temporal adaptive reflectance fusion model (STARFM) and Fit-FC; two typical spatio-temporal fusion methods). The results show that registration error has a significant impact on the accuracy of spatio-temporal fusion: as the registration error increased, the accuracy decreased monotonically. The effect of registration error in a heterogeneous region was greater than that in a homogeneous region. Moreover, the accuracy of fusion was not dependent on the temporal distance between images to be fused, but rather on their statistical correlation. Finally, the Fit-FC method was found to be more accurate than the STARFM method, under all registration error scenarios.

26 August 2022 in IEEE Transactions on Geoscience and Remote Sensing
Verified
Journal Article
Open access
Ka ZhangKa Zhang
Ka Zhang
LYLongjie Ye
Longjie Ye
Wen XIAOWen XIAO
Wen XIAO
YSYehua Sheng
Yehua Sheng
SZShan Zhang
Shan Zhang
YZYaqin Zhou
Yaqin Zhou
With the development of airborne light detection and ranging (LiDAR) technology, it has become a common and efficient way to collect large-scale 3-D spatial information. However, efficient and automatic semantic segmentation of LiDAR data, in the form of 3-D point clouds, remains a persistent challenge. To address this, a dual attention neural network (DA-Net) is proposed, consisting of two different blocks, namely, augmented edge representation (AER) and elevation attentive pooling (EAP). First, the AER can adaptively represent local orientation and position, thereby effectively enhancing geometric information. Second, the captured local features of centroid points are utilized to further encode discriminative features using the EAP with the learned attention scores. Finally, a location homogeneity (LH) module is devised to explore the long-range relationship in an encoder–decoder network. Benefiting from the dual attention module, geometric information hidden in unorganized point clouds can be effectively propagated. Besides, the LH forces the network to pay attention to the semantic consistency of elevated objects, which facilitates both point- and object-level point cloud semantic segmentation for scene understanding. A benchmark dataset is used to assess the proposed method, which achieves an overall accuracy of 85.98% and an average F1 score of 72.31%. In addition, comparisons with other latest deep learning methods on the 2019 Data Fusion Contest dataset further demonstrate the robustness and generalization ability of the proposed method.

1 September 2012 in IET Intelligent Transport Systems
Verified
Journal Article
Closed access
Ka ZhangKa Zhang
Ka Zhang
YSY. Sheng
Y. Sheng
JLJ. Li
J. Li
Considering the problem of automatic information acquisition in the field of intelligent transportation system (ITS), a new approach for detection of road traffic sign from natural scene images is proposed in this study. The adaptive colour segmentation based on pixel vector is firstly used to segment colour image into binary image and stand out traffic sign regions, which can reduce the influence of lighting conditions on image segmentation. Secondly, to improve the ability of shape identification during traffic sign detection, central projection transformation (CPT) is used to compute shape feature vectors of different candidate regions, and this shape feature is input to the probabilistic neural networks (PNN) to discriminate true traffic signs from candidates. The proposed approach is applied to many natural images. Experimental results show that the proposed method can effectively detect road traffic signs from natural scene images.

4 March 2025 in Fire
Verified
Journal Article
Open access
PDF file
CZChengmeng Zhao
Chengmeng Zhao
LZLike Zhao
Like Zhao
Ka ZhangKa Zhang
Ka Zhang
YRYinghua Ren
Yinghua Ren
HCHui Chen
Hui Chen
YSYehua Sheng
Yehua Sheng
Owing to the demand for smoke and flame detection in natural scenes, this paper proposes a lightweight deep learning model, SF-YOLO (Smoke and Fire-YOLO), for video smoke and flame detection in such environments. Firstly, YOLOv11 is employed as the backbone network, combined with the C3k2 module based on a two-path residual attention mechanism, and a target detection head frame with an embedded attention mechanism. This combination enhances the response of the unobscured regions to compensate for the feature loss in occluded regions, thereby addressing the occlusion problem in dynamic backgrounds. Then, a two-channel loss function (W-SIoU) based on dynamic tuning and intelligent focusing is designed to enhance loss computation in the boundary regions, thus improving the YOLOv11 model’s ability to recognize targets with ambiguous boundaries. Finally, the algorithms proposed in this paper are experimentally validated using the self-generated dataset S-Firedata and the public smoke and flame virtual dataset M4SFWD. These datasets are derived from internet smoke and flame video frame extraction images and open-source smoke and flame dataset images, respectively. The experimental results demonstrate, compared with deep learning models such as YOLOv8, Gold-YOLO, and Faster-RCNN, the SF-YOLO model proposed in this paper is more lightweight and exhibits higher detection accuracy and robustness. The metrics mAP50 and mAP50-95 are improved by 2.5% and 2.4%, respectively, in the self-made dataset S-Firedata, and by 0.7% and 1.4%, respectively, in the publicly available dataset M4SFWD. The research presented in this paper provides practical methods for the automatic detection of smoke and flame in natural scenes, which can further enhance the effectiveness of fire monitoring systems.