The recent explosive interest in the reasoning capabilities of large language models, such as DeepSeek-R1, has demonstrated remarkable success through reinforcement learning-based fine-tuning frameworks, exemplified by methods like Group Relative Policy Optimization (GRPO). However, such reasoning abilities remain underexplored and notably absent in vision foundation models, including representation models like the DINO series. In this work, we propose DINO-R1, the first such attempt to incentivize visual in-context reasoning capabilities of vision foundation models using reinforcement learning. Specifically, DINO-R1 introduces Group Relative Query Optimization (GRQO), a novel reinforcement-style training strategy explicitly designed for query-based representation models, which computes query-level rewards based on group-normalized alignment quality. We also apply KL-regularization to stabilize the objectness distribution to reduce the training instability. This joint optimization enables dense and expressive supervision across queries while mitigating overfitting and distributional drift. Building upon Grounding-DINO, we train a series of DINO-R1 family models that integrate a visual prompt encoder and a visual-guided query selection mechanism. Extensive experiments on COCO, LVIS, and ODinW demonstrate that DINO-R1 significantly outperforms supervised fine-tuning baselines, achieving strong generalization in both open-vocabulary and closed-set visual prompting scenarios.

World model based reinforcement learning (RL) has emerged as a promising approach for autonomous driving, which learns a latent dynamics model and uses it to train a planning policy. To speed up the learning process, the pretrain-finetune paradigm is often used, where online RL is initialized by a pretrained model and a policy learned offline. However, naively performing such initialization in RL may result in dramatic performance degradation during the online interactions in the new task. To tackle this challenge, we first analyze the performance degradation and identify two primary root causes therein: the mismatch of the planning policy and the mismatch of the dynamics model, due to distribution shift. We further analyze the effects of these factors on performance degradation during finetuning, and our findings reveal that the choice of finetuning strategies plays a pivotal role in mitigating these effects. We then introduce AdaWM, an Adaptive World Model based planning method, featuring two key steps: (a) mismatch identification, which quantifies the mismatches and informs the finetuning strategy, and (b) alignment-driven finetuning, which selectively updates either the policy or the model as needed using efficient low-rank updates. Extensive experiments on the challenging CARLA driving tasks demonstrate that AdaWM significantly improves the finetuning process, resulting in more robust and efficient performance in autonomous driving systems.

Autonomous driving is a complex and challenging task that aims at safe motion planning through scene understanding and reasoning. While vision-only autonomous driving methods have recently achieved notable performance, through enhanced scene understanding, several key issues, including lack of reasoning, low generalization performance and long-tail scenarios, still need to be addressed. In this paper, we present VLP, a novel Vision-Language-Planning framework that exploits language models to bridge the gap between linguistic understanding and autonomous driving. VLP enhances autonomous driving systems by strengthening both the source memory foundation and the self-driving car's contextual understanding. VLP achieves state-of-the-art end-to-end planning performance on the challenging NuScenes dataset by achieving 35.9% and 60.5% reduction in terms of average L2 error and collision rates, respectively, compared to the previous best method. Moreover, VLP shows improved performance in challenging long-tail scenarios and strong generalization capabilities when faced with new urban environments.

Autonomous driving stands as a pivotal domain in computer vision, shaping the future of transportation. Within this paradigm, the backbone of the system plays a crucial role in interpreting the complex environment. However, a notable challenge has been the loss of clear supervision when it comes to Bird's Eye View elements. To address this limitation, we introduce CLIP-BEVFormer, a novel approach that leverages the power of contrastive learning techniques to enhance the multi-view image-derived BEV backbones with ground truth information flow. We conduct extensive experiments on the challenging nuScenes dataset and showcase significant and consistent improvements over the SOTA. Specifically, CLIP-BEVFormer achieves an impressive 8.5% and 9.2% enhancement in terms of NDS and mAP, respectively, over the previous best BEV model on the 3D object detection task.

Whether by processing videos with fixed resolution from start to end or incorporating pooling and down-scaling strategies, existing video transformers process the whole video content throughout the network without specially handling the large portions of redundant information. In this paper, we present a Supertoken Video Transformer (SVT) that incorporates a Semantic Pooling Module (SPM) to aggregate latent representations along the depth of visual transformer based on their semantics, and thus, reduces redundancy inherent in video inputs.~Qualitative results show that our method can effectively reduce redundancy by merging latent representations with similar semantics and thus increase the proportion of salient information for downstream tasks.~Quantitatively, our method improves the performance of both ViT and MViT while requiring significantly less computations on the Kinectics and Something-Something-V2 benchmarks.~More specifically, with our SPM, we improve the accuracy of MAE-pretrained ViT-B and ViT-L by 1.5% with 33% less GFLOPs and by 0.2% with 55% less FLOPs, respectively, on the Kinectics-400 benchmark, and improve the accuracy of MViTv2-B by 0.2% and 0.3% with 22% less GFLOPs on Kinectics-400 and Something-Something-V2, respectively.

Capturing interaction of hands with objects is important to autonomously detect human actions from egocentric videos. In this work, we present a pyramid video transformer with a dynamic class token generator for egocentric action recognition. Different from previous video transformers, which use the same static embedding as the class token for diverse inputs, we propose a dynamic class token generator that produces a class token for each input video by analyzing the hand-object interaction and the related motion information. The dynamic class token can diffuse such information to the entire model by communicating with other informative tokens in the subsequent transformer layers. With the dynamic class token, dissimilarity between videos can be more prominent, which helps the model distinguish various inputs. In addition, traditional video transformers explore temporal features globally, which requires large amounts of computation. However, egocentric videos often have a large amount of background scene transition, which causes discontinuities across distant frames. In this case, blindly reducing the temporal sampling rate will risk losing crucial information. Hence, we also propose a pyramid architecture to hierarchically process the video from short-term high rate to long-term low rate. With the proposed architecture, we significantly reduce the computational cost as well as the memory requirement without sacrificing from the model performance. We perform comparisons with different baseline video transformers on the EPIC-KITCHENS-100 and EGTEA Gaze+ datasets. Both quantitative and qualitative results show that the proposed model can efficiently improve the performance for egocentric action recognition.

Thermography technology is widely used to inspect thermal anomalies in building façade systems. Computer vision-based techniques provide opportunities to autonomously detect such heat anomalies to significantly improve the efficiency of decision-making for building envelope retrofitting and maintenance. In this work, we propose a novel Capsule Network-based deep learning model – CapsLab – that detects and identifies thermal anomalies by semantic segmentation. CapsLab is built based on our proposed prediction-tuning capsule (PT-Capsule) layer. Different from a traditional capsule layer, which consists of part-whole transformation and capsule-routing process, the proposed layer is composed of a prediction and tuning process, which helps decreasing the number of model parameters significantly. While the applicability of traditional Capsule Networks (CapsNets) has been limited to simpler tasks and smaller datasets due to their scalability issue, we can leverage the lightweight of the proposed PT-Capsule layer, and apply it to the semantic segmentation task. In this work, we also employ our previously presented performance metric, referred to as the Anomaly Identification Metric (AIM) (Kakillioglua et al. 2021), to evaluate the segmentation outputs. Traditional performance metrics do not accurately reflect the true performance of the segmentation models in thermal anomaly identification due to the high subjectivity in the annotation process and higher overlap ratio sensitivity of the standard metrics. AIM, on the other hand, is robust to these drawbacks. Experimental results show, both qualitatively and quantitatively, that our proposed segmentation method can effectively segment the thermal anomalies. Specifically, our model provides 9.38% and 13.53% improvements over the baseline model – DeepLabV3+ – based on traditional mIoU score and the AIM score, respectively, while requiring less model parameters and less computation at the same time. In addition, the scores that the AIM metric generates better align with the scores provided by building performance experts.

Building facades play a significant role in the thermal and energy performance of a building. The presence of thermal anomalies in facades can cause uncontrolled heat leakages, which impacts a facade's thermal performance. Unmanned aerial vehicles (UAVs) equipped with infrared (IR) cameras are becoming increasingly popular as non-destructive methods for the inspection of thermal anomalies in building envelopes. However, the thermal anomalies identified in UAV-captured images are difficult to localize on a 3D building model, which impedes the assessment and documentation of the thermal anomalies with their properties. This paper proposes a computational workflow to automatically register the detected imagery thermal anomalies to a 3D building model. First, IR images and thermal anomalies are sequentially aligned with the corresponding visual RGB images and mapped into facade reference images (RIs) by using image matching techniques. They are then registered to the 3D building coordinate system through a predefined coordinate transformation process. A pilot case study is presented to demonstrate the processing of UAV-images using the proposed method. As a result of this study, the proposed method enables automated registration and data fusion of UAV-captured images, as well as thermal anomaly information into a 3D building model, supporting the assessment of anomalies and evaluation of existing facade conditions.

Capsule Networks (CapsNets) create internal representations by parsing inputs into various instances at different resolution levels via a two-phase process – part-whole transformation and hierarchical component routing. Since both of these internal phases are computationally expensive, CapsNet have not found wider use. Existing variations of CapsNets mainly focus on performance comparison with the original CapsNet, and have not outperformed CNN-based models on complex tasks. To address the limitations of the existing CapsNet structures, we propose a novel Prediction-Tuning Capsule Network (PT-CapsNet), and also introduce fully connected PT-Capsules (FC-PT-Caps) and locally connected PT-Capsules (LC-PT-Caps). Different from existing CapsNet structures, our proposed model (i) allows the use of capsules for more difficult vision tasks and provides wider applicability; and (ii) provides better than or comparable performance to CNN-based baselines on these complex tasks. In our experiments, we show robustness to affine transformations, as well as the lightweight and scalability of PT-CapsNet via constructing larger and deeper networks and performing comparisons on classification, semantic segmentation and object detection tasks. The results show consistent performance improvement and significant parameter reduction compared to various baseline models. Code is available at https://github.com/Christinepan881/PT-CapsNet.git.