實驗室研究面向自動攝影的具身智能體:

1、感知:
(1)深度視覺計算:深度視覺計算是利用計算機視覺技術估計圖像或視頻中的深度信息,即場景中各點像素到相機成像平面的垂直距離。實驗室關注室內外場景無縫切換的跨域單目度量深度估計。

(2)開放世界感知與導航:①開放世界2D或3D障礙物/未知物體檢測與識別;②開放世界視覺導航:視覺導航是移動機器人利用視覺傳感器實現場景感知、路徑規劃、運動規劃的整個體系。實驗室關注視覺導航及避障技術。包括: 視覺里程計(VO)、建圖(利用VO和深度圖)、重定位(從已知地圖中識別自身位置)、閉環檢測(消除VO的閉環誤差) 、障礙物檢測、道路分割、未知物體檢測、全局導航、視覺避障、Scene tagging(自動標注房間中物體)等。

2、推理:根據已有的知識和信息進行邏輯思考,做出合理的判斷。
(1)開放世界的數字孿生;
(2)知識與數據雙輪驅動;
3、決策。基于推理的結果選擇最佳的行動方案。
4、記憶。短期工作記憶和三種長期記憶:情景記憶、語義記憶和程序記憶。
vRobotit實驗室關于“具身攝影智能體”代表性論文:
[1] Yihao Liu, Feng Xue,, Anlong Ming*, Mingshuai Zhao, Huadong Ma, Nicu Sebe, SM4Depth: Seamless Monocular Metric Depth Estimation across Multiple Cameras and Scenes by One Model, in Proceedings of the 32th ACM International Conference on Multimedia (MM), 2024. 注:深度視覺計算
代碼:arXiv:2403.08556
[2] Feng Xue, Yicong Chang, Tianxi Wang, Yu Zhou, Anlong Ming, Indoor Obstacle Discovery on Reflective Ground Using Monocular Camera, International Journal of Computer Vision (IJCV), vol. 132, pp. 987-1007, 2024. 注:小障礙物感知
代碼:https://github.com/mRobotit/IndoorObstacleDiscovery-RG
[3] Fei Sheng, Feng Xue, Wenteng Liang, Yichong Chang, Anlong Ming*, Monocular Depth Distribution Alignment with Low Computation,the 2022 International Conference on Robotics and Automation (ICRA), 2022. 注:深度視覺計算
代碼:https://github.com/mRobotit/USNet
[4] Feng Xue, Junfeng Cao, Fei Sheng, Yankai Wang, Yu Zhou, Anlong Ming, Boundary-induced and Scene-aggregated Network for Monocular Depth Prediction, Pattern Recognition (PR), vol. 115, 2021. 注:深度視覺計算
代碼:https://github.com/mRobotit/BS-Net
[5] Wenteng Liang, Feng Xue, Yihao Liu, Guofeng Zhong, Anlong Ming*, Unknown Sniffer for Object Detection: Don't Turn a Blind Eye to Unknown Objects, IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2023. 注:開放世界物體感知
代碼:https://github.com/mRobotit/UnSniffer
[6] Y. Chang, F. Xue, F. Sheng, W. Liang and A. Ming, Fast Road Segmentation via Uncertainty-aware Symmetric Network, in International Conference on Robotics and Automation (ICRA), 2022. 注:可行域分割
代碼:https://github.com/mRobotit/DANet
[7] F. Xue, A. Ming and Y. Zhou, Tiny Obstacle Discovery by Occlusion-Aware Multilayer Regression, in IEEE Transactions on Image Processing (TIP), vol. 29, pp. 9373-9386, 2020. 注:小障礙物感知
代碼:https://github.com/mRobotit/Tiny-Obstacle-Discovery-ROS
[8] F. Xue, A. Ming, M. Zhou and Y. Zhou, A Novel Multilayer Framework for Tiny Obstacle Discovery, in International Conference on Robotics and Automation (ICRA), 2019. 注:小障礙物感知
代碼:https://github.com/mRobotit/Tiny-Obstacle-Discovery
[9] Feng Xue, Yicong Chang, Wenzhuang Xu, Wenteng Liang, Fei Sheng, Anlong Ming*, Evidence-based Real-time Road Segmentation with RGB-D Data Augmentation, IEEE Transactions on Intelligent Transportation Systems (TITS), accepted, 2024. 注:可行域分割
代碼:整理中