Xin Ma

Geigh Zollicoffer, Tanush Chopra, Mingkuan Yan et al.

AI systems deployed in the real world must contend with distractions and out-of-distribution (OOD) noise that can destabilize their policies and lead to unsafe behavior. While robust training can reduce sensitivity to some forms of noise, it is infeasible to anticipate all possible OOD conditions. To mitigate this issue, we develop an algorithm that leverages a world model's inherent measure of surprise to reduce the impact of noise in world model--based reinforcement learning agents. We introduce both multi-representation and single-representation rejection sampling, enabling robustness to settings with multiple faulty sensors or a single faulty sensor. While the introduction of noise typically degrades agent performance, we show that our techniques preserve performance relative to baselines under varying types and levels of noise across multiple environments within self-driving simulation domains (CARLA and Safety Gymnasium). Furthermore, we demonstrate that our methods enhance the stability of two state-of-the-art world models with markedly different underlying architectures: Cosmos and DreamerV3. Together, these results highlight the robustness of our approach across world modeling domains. We release our code at https://github.com/Bluefin-Tuna/WISER .

Chuanxin Song, Xin Ma

Despite the remarkable success of convolutional neural networks in various computer vision tasks, recognizing indoor scenes still presents a significant challenge due to their complex composition. Consequently, effectively leveraging semantic information in the scene has been a key issue in advancing indoor scene recognition. Unfortunately, the accuracy of semantic segmentation has limited the effectiveness of existing approaches for leveraging semantic information. As a result, many of these approaches remain at the stage of auxiliary labeling or co-occurrence statistics, with few exploring the contextual relationships between the semantic elements directly within the scene. In this paper, we propose the Semantic Region Relationship Model (SRRM), which starts directly from the semantic information inside the scene. Specifically, SRRM adopts an adaptive and efficient approach to mitigate the negative impact of semantic ambiguity and then models the semantic region relationship to perform scene recognition. Additionally, to more comprehensively exploit the information contained in the scene, we combine the proposed SRRM with the PlacesCNN module to create the Combined Semantic Region Relation Model (CSRRM), and propose a novel information combining approach to effectively explore the complementary contents between them. CSRRM significantly outperforms the SOTA methods on the MIT Indoor 67, reduced Places365 dataset, and SUN RGB-D without retraining. The code is available at: https://github.com/ChuanxinSong/SRRM

Shan An, Guangfu Che, Fangru Zhou et al.

Visual loop closure detection, which can be considered as an image retrieval task, is an important problem in SLAM (Simultaneous Localization and Mapping) systems. The frequently used bag-of-words (BoW) models can achieve high precision and moderate recall. However, the requirement for lower time costs and fewer memory costs for mobile robot applications is not well satisfied. In this paper, we propose a novel loop closure detection framework titled `FILD' (Fast and Incremental Loop closure Detection), which focuses on an on-line and incremental graph vocabulary construction for fast loop closure detection. The global and local features of frames are extracted using the Convolutional Neural Networks (CNN) and SURF on the GPU, which guarantee extremely fast extraction speeds. The graph vocabulary construction is based on one type of proximity graph, named Hierarchical Navigable Small World (HNSW) graphs, which is modified to adapt to this specific application. In addition, this process is coupled with a novel strategy for real-time geometrical verification, which only keeps binary hash codes and significantly saves on memory usage. Extensive experiments on several publicly available datasets show that the proposed approach can achieve fairly good recall at 100\% precision compared to other state-of-the-art methods. The source code can be downloaded at https://github.com/AnshanTJU/FILD for further studies.

Xiaochen Huang, Xiaochen Bi, Cuihua Lv et al.

Parkinson's disease (PD), a prevalent neurodegenerative disorder, significantly affects patients' daily functioning and social interactions. To facilitate a more efficient and accessible diagnostic approach for PD, we propose a dynamic facial expression analysis-based PD auxiliary diagnosis method. This method targets hypomimia, a characteristic clinical symptom of PD, by analyzing two manifestations: reduced facial expressivity and facial rigidity, thereby facilitating the diagnosis process. We develop a multimodal facial expression analysis network to extract expression intensity features during patients' performance of various facial expressions. This network leverages the CLIP architecture to integrate visual and textual features while preserving the temporal dynamics of facial expressions. Subsequently, the expression intensity features are processed and input into an LSTM-based classification network for PD diagnosis. Our method achieves an accuracy of 93.1%, outperforming other in-vitro PD diagnostic approaches. This technique offers a more convenient detection method for potential PD patients, improving their diagnostic experience.

Xin Ma, Xinwu Liang

Most of the state-of-the-art indirect visual SLAM methods are based on the sparse point features. However, it is hard to find enough reliable point features for state estimation in the case of low-textured scenes. Line features are abundant in urban and indoor scenes. Recent studies have shown that the combination of point and line features can provide better accuracy despite the decrease in computational efficiency. In this paper, measurements of point and line features are extracted from RGB-D data to create map features, and points on a line are treated as keypoints. We propose an extended approach to make more use of line observation information. And we prove that, in the local bundle adjustment, the estimation uncertainty of keyframe poses can be reduced when considering more landmarks with independent measurements in the optimization process. Experimental results on two public RGB-D datasets demonstrate that the proposed method has better robustness and accuracy in challenging environments.

Abbaraju Praveen, Xin Ma, Harikrishnan Manoj et al.

Inspection for structural properties (surface stiffness and coefficient of restitution) is crucial for understanding and performing aerial manipulations in unknown environments, with little to no prior knowledge on their state. Inspection-on-the-fly is the uncanny ability of humans to infer states during manipulation, reducing the necessity to perform inspection and manipulation separately. This paper presents an infrastructure for inspection-on-the-fly method for aerial manipulators using hybrid physical interaction control. With the proposed method, structural properties (surface stiffness and coefficient of restitution) can be estimated during physical interactions. A three-stage hybrid physical interaction control paradigm is presented to robustly approach, acquire and impart a desired force signature onto a surface. This is achieved by combining a hybrid force/motion controller with a model-based feed-forward impact control as intermediate phase. The proposed controller ensures a steady transition from unconstrained motion control to constrained force control, while reducing the lag associated with the force control phase. And an underlying Operational Space dynamic configuration manager permits complex, redundant vehicle/arm combinations. Experiments were carried out in a mock-up of a Dept. of Energy exhaust shaft, to show the effectiveness of the inspection-on-the-fly method to determine the structural properties of the target surface and the performance of the hybrid physical interaction controller in reducing the lag associated with force control phase.