Thermal object detection is a critical task in various fields, such as surveillance and autonomous driving. Current state-of-the-art (SOTA) models always leverage a prior Thermal-To-Visible (T2V) translation model to obtain visible spectrum information, followed by a cross-modality aggregation module to fuse information from both modalities. However, this fusion approach does not fully exploit the complementary visible spectrum information beneficial for thermal detection. To address this issue, we propose a novel cross-modal fusion method called Pseudo Visible Feature Fine-Grained Fusion (PFGF). Specifically, a graph is constructed with nodes generated from multi-level thermal features and pseudo-visual latent features produced by the T2V model. Each level of features corresponds to a subgraph. An Inter-Mamba block is proposed to perform cross-modality fusion between nodes at the lowest level; while a Cascade Knowledge Integration (CKI) strategy is used to fuse low-level fused information to high-level subgraphs in a cascade manner. After several iterations of graph node updating, each subgraph outputs an aggregated feature to the detection head respectively. Unlike previous cross-modal fusion methods, our approach explicitly models high-level relationships between cross-modal data, effectively fusing different granularity information. Experimental results demonstrate that our method achieves SOTA detection performance. Code is available at https://github.com/liting1018/PFGF.

Test-time adaptation with pre-trained vision-language models, such as CLIP, aims to adapt the model to new, potentially out-of-distribution test data. Existing methods calculate the similarity between visual embedding and learnable class embeddings, which are initialized by text embeddings, for zero-shot image classification. In this work, we first analyze this process based on Bayes theorem, and observe that the core factors influencing the final prediction are the likelihood and the prior. However, existing methods essentially focus on adapting class embeddings to adapt likelihood, but they often ignore the importance of prior. To address this gap, we propose a novel approach, Bayesian Class Adaptation (BCA), which in addition to continuously updating class embeddings to adapt likelihood, also uses the posterior of incoming samples to continuously update the prior for each class embedding. This dual updating mechanism allows the model to better adapt to distribution shifts and achieve higher prediction accuracy. Our method not only surpasses existing approaches in terms of performance metrics but also maintains superior inference rates and memory usage, making it highly efficient and practical for real-world applications.

Visible-infrared object detection has gained significant attention because of its applications in autonomous driving, video surveillance, and related fields. The effective fusion of multimodal information is fundamental to its success. The existing approaches concentrate on improving the pixel-level fusion mechanisms; detection performance has reached a plateau. We propose a new framework for SAM-guided semantic knowledge fusion ( SemFusion ). The core idea is to leverage semantic priors from large models while incorporating a lightweight cross-modal fusion strategy. Specifically, our method comprises two stages. In the first stage, the Flow-Guided RGB Feature Alignment (FGRA) module establishes object-aware correspondences between multimodalities based on SAM-generated masks. This ensures semantic-level feature matching by deformable convolution alignment. In the second stage, the Semantic Knowledge Distillation (SKD) strategy facilitates the transfer of large-model knowledge to the detection model through SAM feature, offset, and mask level distillations. For the detector model, three blocks are designed to augment any off-the-shelf detector. They are deformable cross-modal alignment, spatio-channel preliminary fusion, and mask-guided feature refinement. By alignment with SAM masks, semantic alignment and fusion can be achieved, breaking the pixel-level fusion barrier. Extensive experiments demonstrate that our method, as a plugin, exhibits superior performance on the DroneVehicle, VEDAI, and LLVIP datasets. Code is available at https://github.com/liting1018/SemFusion.

Unmanned Aerial Vehicle (UAV) object detection faces significant challenges due to complex environmental conditions and different imaging conditions. These factors introduce significant changes in scale and appearance, particularly for small objects that occupy limited pixels and exhibit limited information, complicating detection tasks. To address these challenges, we propose a Multimodel Causal Reasoning framework based on YOLO backbone for UAV Object Detection (MCR-UOD). The key idea is to use the backdoor adjustment to discover the condition-invariant object representation for easy detection. Specifically, the YOLO backbone is first adjusted to incorporate the pre-trained vision-language model. The original category labels are replaced with semantic text prompts, and the detection head is replaced with text-image contrastive learning. Based on this backbone, our method consists of two parts. The first part, named language guided region exploration, discovers the regions with high probability of object existence using text embeddings based on vision-language model such as CLIP. Another part is the backdoor adjustment casual reasoning module, which constructs a confounder dictionary tailored to different imaging conditions to capture global image semantics and derives a prior probability distribution of shooting conditions. During causal inference, we use the confounder dictionary and the prior to intervene on local instance features, disentangling condition variations, and obtaining condition-invariant representations. Experimental results on several public datasets confirm the state-of-the-art performance of our approach. The code, data and models will be released upon publication of this paper.

Black-box domain adaptation treats the source domain model as a black box. During the transfer process, the only available information about the target domain is the noisy labels output by the black-box model. This poses significant challenges for domain adaptation. Conventional approaches typically tackle the black-box noisy label problem from two aspects: self-knowledge distillation and pseudo-label denoising, both achieving limited performance due to limited knowledge information. To mitigate this issue, we explore the potential of off-the-shelf vision-language (ViL) multimodal models with rich semantic information for black-box domain adaptation by introducing an Adversarial Experts Model (AEM). Specifically, our target domain model is designed as one feature extractor and two classifiers, trained over two stages: In the knowledge transferring stage, with a shared feature extractor, the black-box source model and the ViL model act as two distinct experts for joint knowledge contribution, guiding the learning of one classifier each. While contributing their respective knowledge, the experts are also updated due to their own limitation and bias. In the adversarial alignment stage, to further distill expert knowledge to the target domain model, adversarial learning is conducted between the feature extractor and the two classifiers. A new consistency-max loss function is proposed to measure two classifier consistency and further improve classifier prediction certainty. Extensive experiments on multiple datasets demonstrate the effectiveness of our approach. Code is available at https://github.com/singinger/AEM.

We propose to explore an interesting and promising problem, Cloud Object Detector Adaptation (CODA), where the target domain leverages detections provided by a large cloud model to build a target detector. Despite with powerful generalization capability, the cloud model still cannot achieve error-free detection in a specific target domain. In this work, we present a novel Cloud Object detector adaptation method by Integrating different source kNowledge (COIN). The key idea is to incorporate a public vision-language model (CLIP) to distill positive knowledge while refining negative knowledge for adaptation by self-promotion gradient direction alignment. To that end, knowledge dissemination, separation, and distillation are carried out successively. Knowledge dissemination combines knowledge from cloud detector and CLIP model to initialize a target detector and a CLIP detector in target domain. By matching CLIP detector with the cloud detector, knowledge separation categorizes detections into three parts: consistent, inconsistent and private detections such that divide-and-conquer strategy can be used for knowledge distillation. Consistent and private detections are directly used to train target detector; while inconsistent detections are fused based on a consistent knowledge generation network, which is trained by aligning the gradient direction of inconsistent detections to that of consistent detections, because it provides a direction toward an optimal target detector. Experiment results demonstrate that the proposed COIN method achieves the state-of-the-art performance.

Unsupervised domain adaptation aims to transfer knowledge from labeled source domain to unlabeled target domain. The semi-supervised method based on mean-teacher framework is one of the main stream approaches. By enforcing consistency constraints, it is hopeful that the teacher network will distill useful source domain knowledge to the student network. However, in practice negative transfer often emerges because the performance of the teacher network is not guaranteed to be always better than the student network. To address this limitation, a novel Adaptive Mutual Learning (AML) strategy is proposed in this paper. Specifically, given a target sample, the network with worse prediction will be optimized by pushing its prediction close to the better prediction. This is in the spirit of traditional knowledge distillation. On the other hand, the network with better prediction is further refined by requiring its prediction to stay away from the worse prediction. This can be regarded conceptually as reverse knowledge distillation. In this way, two networks learn from each other according to their respective performance. At inference phase, the averaged output of these two networks can be taken as the final prediction. Experimental results demonstrate that our AML achieves competitive results.

Source-free object detection (SFOD) needs to adapt a detector pre-trained on a labeled source domain to a tar-get domain, with only unlabeled training data from the tar-get domain. Existing SFOD methods typically adopt the pseudo labeling paradigm with model adaption alternating between predicting pseudo labels and fine-tuning the model. This approach suffers from both unsatisfactory accuracy of pseudo labels due to the presence of domain shift and lim-ited use of target domain training data. In this work, we present a novel Learning to Overlook Domain Style (LODS) method with such limitations solved in a principled man-ner. Our idea is to reduce the domain shift effect by en-forcing the model to overlook the target domain style, such that model adaptation is simplified and becomes easier to carry on. To that end, we enhance the style of each tar-get domain image and leverage the style degree difference between the original image and the enhanced image as a self-supervised signal for model adaptation. By treating the enhanced image as an auxiliary view, we exploit a student- teacher architecture for learning to overlook the style de-gree difference against the original image, also character-ized with a novel style enhancement algorithm and graph alignment constraint. Extensive experiments demonstrate that our LODS yields new state-of-the-art performance on four benchmarks.

As a state-of-the-art family of Unsupervised Domain Adaptation (UDA), bi-classifier adversarial learning methods are formulated in an adversarial (minimax) learning framework with a single feature extractor and two classifiers. Model training alternates between two steps: (I) constraining the learning of the two classifiers to maximize the prediction discrepancy of unlabeled target domain data, and (II) constraining the learning of the feature extractor to minimize this discrepancy. Despite being an elegant formulation, this approach has a fundamental limitation: Maximizing and minimizing the classifier discrepancy is not class discriminative for the target domain, finally leading to a suboptimal adapted model. To solve this problem, we propose a novel Class Discriminative Adversarial Learning (CDAL) method characterized by discovering class discrimination knowledge and leveraging this knowledge to discriminatively regulate the classifier discrepancy constraints on-the-fly. This is realized by introducing an evaluation criterion for judging each classifier's capability and each target domain sample's feature reorientation via objective loss reformulation. Extensive experiments on three standard benchmarks show that our CDAL method yields new state-of-the-art performance. Our code is made available at https://github.com/buerzlh/CDAL.

Unsupervised cross-domain object detection transfers a detection model trained on a source domain to the target domain that has a different data distribution from the source domain. Conventional domain adaptation detection protocols need source domain data during adaptation. However, due to some reasons such as data security, privacy and storage, we cannot access the source data in many practical applications. In this paper, we focus on source-data free domain adaptive object detection, which uses the pre-trained source model instead of the source data for cross-domain adaptation. Due to the lack of source data, we cannot directly align domain distribution between domains. To challenge this, we propose the Source style transferred Mean Teacher (SMT) for source-data free Object Detection. The batch normalization layers in the pre-trained model contain the style information and the data distribution of the non-observed source data. Thus we use the batch normalization information from the pre-trained source model to transfer the target domain feature to the source-like style feature to make full use of the knowledge from the pre-trained source model. Meanwhile, we use the consistent regularization of the Mean Teacher to further distill the knowledge from the source domain to the target domain. Furthermore, we found that by adding perturbations associated with the target domain distribution, the model can increase the robustness of domain-specific information, thus making the learned model generalized to the target domain. Experiments on multiple domain adaptation object detection benchmarks verify that our method is able to achieve state-of-the-art performance.