Can Tian

Dr. Can Tian

orcid

Research Focus & Metrics

Academic Career Length

5Years (2022-Present)

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

Representative Publications

View All Publications
17 January 2023 in Minerals Engineering
Verified
Journal Article
Closed access
CTCan Tian
Can Tian
ZTZhaohui Tang
Zhaohui Tang
YXYongfang Xie
Yongfang Xie
HZHu Zhang
Hu Zhang
WGWeihua Gui
Weihua Gui

1 January 2022 in Powder Technology
Verified
Journal Article
Closed access
CTCan Tian
Can Tian
ZTZhaohui Tang
Zhaohui Tang
HZHu Zhang
Hu Zhang
XGXiaoliang Gao
Xiaoliang Gao
YXYongfang Xie
Yongfang Xie

21 July 2022 in IEEE Transactions on Instrumentation and Measurement
Verified
Journal Article
Closed access
CTCan Tian
Can Tian
ZTZhaohui Tang
Zhaohui Tang
HZHu Zhang
Hu Zhang
YXYongfang Xie
Yongfang Xie
Accurate recognition of operating conditions helps to correctly control froth flotation. Most previous recognition methods mainly focus on the statistic feature but ignore the temporal information in the dynamic and changing flotation process. Actually, the temporal changing trend of froth appearance can reflect the change in flotation performance. Besides, the bubble size distribution (BSD) is a representative visual feature for characterizing froth appearance. Therefore, we propose to exploit the statistics and temporal correlation of BSD for condition recognition in this article. First, to sufficiently characterize the statistics of bubble sizes about the initial state of the froth video, we sample multiple froth images from a short video clip at the beginning of the video and then use a cumulative distribution function (CDF) to describe the BSD in these sampled images, which is regarded as the cumulative distribution statistic (CDS). Meanwhile, to obtain the temporal change trend of BSDs from the initial to the last moment of the froth video, we sample multiple sequential froth images from the whole video and use the CDF to describe the BSD of individual image and then the BSD sequences are input into the bidirectional long short-term memory (BiLSTM) to extract their temporal correlation, which is regarded as the cumulative distribution temporal correlation (CDTC). Additionally, we integrate texture and velocity features with our constructed CDS and CDTC features via a joint learning network to recognize operating conditions. Extensive experiments demonstrate that the proposed model using CDS and CDTC is more effective than other models with BSD features. Besides, compared to the models using the single CDS and CDTC, the accuracy of the proposed model is increased by 8.39% and 2.61%, respectively.

11 April 2025 in Proceedings of the AAAI Conference on Artificial Intelligence
Verified
Conference Paper
Closed access
PDF file
YWYunbo Wang
Yunbo Wang
YWYuJie Wu
YuJie Wu
ZDZhien Dai
Zhien Dai
CTCan Tian
Can Tian
JLJun Long
Jun Long
JCJianhai Chen
Jianhai Chen
Cross-modal matching shows enormous potential to recognize objects across different sensory modalities, which is fundamental to numerous visual-language tasks like image-text retrieval and visual captioning. Existing works generally rely on massive and well-aligned data pairs for model training. Unfortunately, multimodal datasets are extremely difficult to annotate and collect. As an alternative, the co-occurred data pairs collected from the internet have been widely exploited to train a cross-modal matching model. However, the cheaply-collected dataset unavoidably contains mismatched pairs (i.e., noisy correspondence), which are detrimental to the matching model. In this paper, we propose an alternative method termed noisy correspondence rectification via Asymmetric Similarity Learning (ASL), and it allows for dealing with insufficient learning of positive and negative pairs caused by the popular triplet-based symmetric learning fashion. Specifically, the learning of positive or negative pairs within a triplet is conducted in an asymmetric fashion, and the self-paced weighting boundary is imposed on positive pairs to mitigate the effect of noise. Meanwhile, the optimization of negative samples will not be affected in the process of punishing potentially-noisy positive samples. To verify the effectiveness of our proposed approach, a series of experiments are conducted on three widely-used benchmarks (i.e., Flick30k, MS-COCO and CC152k), and the results show superior performance compared to the state-of-the-art methods.

21 February 2024 in Control Engineering Practice
Verified
Journal Article
Closed access
CTCan Tian
Can Tian
ZTZhaohui Tang
Zhaohui Tang
HZHu Zhang
Hu Zhang
YXYongfang Xie
Yongfang Xie
ZDZhien Dai
Zhien Dai

28 October 2024
Verified
Conference Paper
Open access
ZDZhien Dai
Zhien Dai
ZTZhaohui Tang
Zhaohui Tang
HZHu Zhang
Hu Zhang
CTCan Tian
Can Tian
MPMingjun Pan
Mingjun Pan
YXYongfang Xie
Yongfang Xie
Stereo matching is a pivotal technique for depth estimation and has been popularly applied in various computer vision tasks. Although many related methods have been reported recently, they still face some challenges such as significant disparity variations at object boundaries, difficult prediction at large disparity regions, and suboptimal generalization when label distribution varies between source and target domains. Therefore, we propose a stereo-matching model (i.e., Eglcr) that utilizes edge structure information and multi-scale matching similarity features for better disparity estimation. First, we use a lightweight network to predict the initial disparity. Then, we develop a multi-scale similarity feature extraction module, incorporating adaptive attention mechanisms, to capture the fusion similarity information of stereo images across various scales. Meanwhile, we introduce an edge structure-aware module that features an iteratively optimized disparity map and a scale attention factor, aimed at accurately delineating edge information in complex scenes. After that, we employ an iterative strategy for disparity estimation, guided by the fusion similarity features across multiple scales and the detailed edge structure information. We conduct abundant experiments on some popular stereo matching datasets including Middlebury, KITTI, ETH3D, and Scene Flow. The results show that our proposed Eglcr achieves state-of-the-art performance both in accuracy and generalization. Our code is available at https://github.com/kangarooCV/Eglcr.