TEOChat: A Large Vision-Language Assistant for Temporal Earth Observation Data
Jeremy Andrew Irvin, Emily Ruoyu Liu, Joyce Chuyi Chen, Ines Dormoy, Jinyoung Kim, Samar Khanna, Zhuo Zheng, Stefano Ermom
International Conference on Learning Representations (ICLR), 2025
paper    code     TEOChatlas dataset     TEOChat demo    
Highlight: TEOChatlas, the first temporal EO instruction-following dataset that has >500k instruction-following examples, spanning dozens of spatial and temporal reasoning tasks. TEOChat is the first VLM that can converse about temporal earth observation imagery.

Changen2: Multi-Temporal Remote Sensing Generative Change Foundation Model
Zhuo Zheng, Stefano Ermon, Dongjun Kim, Liangpei Zhang, Yanfei Zhong
IEEE Transactions on Pattern Analysis and Machine Intelligence (TPAMI), 2024
SCI Q1 Top
paper    arxiv    code     Changen2-S1-15k dataset     Changen2-S9-27k dataset
Highlight: Generative Change Foundation Model, Changen2 can be trained at scale using self-supervision, yielding change supervisory signals from unlabeled single-temporal images; The resulting task-specific foundation model possesses inherent zero-shot change detection capabilities and excellent transferability.

Segment Any Change
Zhuo Zheng, Yanfei Zhong, Liangpei Zhang, Stefano Ermon
38th Annual Conference on Neural Information Processing Systems (NeurIPS), 2024
paper    code
Highlight: Zero-shot Change Detection is introduced for the first time; Training-free Adaptation for Foundation models.

Towards transferable building damage assessment via unsupervised single-temporal change adaptation
Zhuo Zheng, Yanfei Zhong, Liangpei Zhang, Marshall Burke, David B. Lobell, Stefano Ermon
Remote Sensing of Environment (RSE), 2024
SCI Q1 Top, ranking it 1 out of 32 in Remote Sensing
paper   
Highlight: Transferable building damage assessment; Unsupervised single-temporal change adaptation (STCA) enables models to achieve adaptation with only target pre-disaster images.

Unifying Remote Sensing Change Detection via Deep Probabilistic Change Models: From Principles, Models to Applications
Zhuo Zheng, Yanfei Zhong, Ji Zhao, Ailong Ma, Liangpei Zhang
ISPRS Journal of Photogrammetry and Remote Sensing (ISPRS P&RS), 2024
SCI Q1 Top, ranking it 1 out of 50 in Geography, Physical
paper    code
Highlight: Unified probabilistic change modeling: Probabilistic Change Model (PCM); and strong model instance: ChangeSparse.

Single-Temporal Supervised Learning for Universal Remote Sensing Change Detection
Zhuo Zheng, Ailong Ma, Liangpei Zhang, Yanfei Zhong
International Journal of Computer Vision (IJCV), 2024
SCI Q1 Top
paper    code
Highlight: ChangeStar2: STAR with Faster and More Stable Convergence; Universal Remote Sensing Change Detection (object, semantic, class-agnostic, time-series change)

Scalable Multi-Temporal Remote Sensing Change Data Generation via Simulating Stochastic Change Process
Zhuo Zheng, Shiqi Tian, Ailong Ma, Liangpei Zhang, Yanfei Zhong
International Conference on Computer Vision (ICCV), 2023
paper    code
Highlight: New direction: Generative Change Modeling; Changen, Change generator, enables object change generation with controllable object property and change event; Effective synthetic change data pre-training.

FarSeg++: Foreground-Aware Relation Network for Geospatial Object Segmentation in High Spatial Resolution Remote Sensing Imagery
Zhuo Zheng, Yanfei Zhong, Junjue Wang, Ailong Ma, Liangpei Zhang
IEEE Transactions on Pattern Analysis and Machine Intelligence (TPAMI), 2023
SCI Q1 Top
paper    code    UV6K dataset

ChangeMask: Deep Multi-task Encoder-Transformer-Decoder Architecture for Semantic Change Detection
Zhuo Zheng, Yanfei Zhong, Shiqi Tian, Ailong Ma, Liangpei Zhang
ISPRS Journal of Photogrammetry and Remote Sensing (ISPRS P&RS), 2022
SCI Q1 Top, ranking it 1 out of 50 in Geography, Physical
paper     code     SECOND dataset: train / val split (ours)
Highlight: Disentangled semantic and change representation and temporal symmetric transformer (TST) for semantic change detection.
ESI Highly Cited Paper, ISPRS P&RS Top Cited Paper

Building Damage Assessment for Rapid Disaster Response with a Deep Object-based Semantic Change Detection Framework: from natural disasters to man-made disasters
Zhuo Zheng, Yanfei Zhong, Junjue Wang, Ailong Ma, Liangpei Zhang
Remote Sensing of Environment (RSE), 2021
SCI Q1 Top, ranking it 1 out of 32 in Remote Sensing
paper    code
Highlight: A deep object-based semantic change detection method for building damage assessment in the context of disaster response.
ESI Highly Cited Paper, RSE Top Cited Paper
Our solution is xView2 4th overall officially, which is a unique single-model solution among top-5.

Change is Everywhere: Single-Temporal Supervised Object Change Detection in Remote Sensing Imagery
Zhuo Zheng, Ailong Ma, Liangpei Zhang, Yanfei Zhong
International Conference on Computer Vision (ICCV), 2021
paper     project     code
Highlight: A single-temporal supervised learning algorithm and a plug-and-play change detection head: ChangeMixin for change detection.
The method has been included in microsoft/torchgeo and PaddlePaddle/PaddleRS.

LoveDA: A Remote Sensing Land-Cover Dataset for Domain Adaptive Semantic Segmentation
Junjue Wang^*, Zhuo Zheng^*, Ailong Ma, Xiaoyan Lu, Yanfei Zhong (* Equal contribution)
35th Annual Conference on Neural Information Processing Systems (NeurIPS Datasets & Benchmarks Track), 2021
paper     dataset/code

Deep Multisensor Learning for Missing-Modality All-Weather Mapping
Zhuo Zheng, Ailong Ma, Liangpei Zhang, Yanfei Zhong
ISPRS Journal of Photogrammetry and Remote Sensing (ISPRS P&RS), 2021
SCI Q1 Top, ranking it 1 out of 50 in Geography, Physical
paper
Highlight: A registration-free multi-modal/sensor learning algorithm via exploring meta-modal/sensory representation for all-weather mapping.

Foreground-Aware Relation Network for Geospatial Object Segmentation in High Spatial Resolution Remote Sensing Imagery
Zhuo Zheng, Yanfei Zhong, Junjue Wang, Ailong Ma
Computer Vision and Pattern Recognition (CVPR), 2020
paper     code
Highlight: Explicit foreground modeling from the perspectives of relation and optimization for real-time geospatial object segmentation.
The method has been included in microsoft/torchgeo and PaddlePaddle/PaddleRS.

FPGA: Fast Patch-Free Global Learning Framework for Fully End-to-End Hyperspectral Image Classification
Zhuo Zheng, Yanfei Zhong, Ailong Ma, Liangpei Zhang
IEEE Transactions on Geoscience and Remote Sensing (TGRS), 2020
SCI Q1 Top
paper     code
Highlight: Patch-free is all you need towards faster and stronger hyperspectral image classification.
The method has been included in WHULuoJiaTeam/luojianet.
ESI Highly Cited Paper

HyNet: Hyper-scale object detection network framework for multiple spatial resolution remote sensing imagery
Zhuo Zheng, Yanfei Zhong, Ailong Ma, Xiaobing Han, Ji Zhao, Yanfei Liu, Liangpei Zhang
ISPRS Journal of Photogrammetry and Remote Sensing (ISPRS P&RS), 2020
SCI Q1 Top, ranking it 1 out of 50 in Geography, Physical
paper
Highlight: Hyper-scale = Multi-scale × Multi-scale, a new perspective of scale modeling at the convolutional groups.

COLOR: Cycling, Offline Learning, and Online Representation Framework for Airport and Airplane Detection Using GF-2 Satellite Images
Yanfei Zhong, Zhuo Zheng^*, Ailong Ma, Xiaoyan Lu, Liangpei Zhang   (* denotes the corresponding author)
IEEE Transactions on Geoscience and Remote Sensing (TGRS), 2020
SCI Q1 Top
paper
Highlight: Evolvable model and dataset: making your model and dataset both great again for new domain data.