This paper introduces reconstructive visual instruction tuning (ROSS), a family of Large Multimodal Models (LMMs) that exploit vision-centric supervision signals. In contrast to conventional visual instruction tuning approaches that exclusively supervise text outputs, ROSS prompts LMMs to supervise visual outputs via reconstructing input images. By doing so, it capitalizes on the inherent richness and detail present within input images themselves, which are often lost in pure text supervision. However, producing meaningful feedback from natural images is challenging due to the heavy spatial redundancy of visual signals. To address this issue, ROSS employs a denoising objective to reconstruct latent representations of input images, avoiding directly regressing exact raw RGB values. This intrinsic activation design inherently encourages LMMs to maintain image detail, thereby enhancing their fine-grained comprehension capabilities and reducing hallucinations. Empirically, ROSS consistently brings significant improvements across different visual encoders and language models. In comparison with extrinsic assistance state-of-the-art alternatives that aggregate multiple visual experts, ROSS delivers competitive performance with a single SigLIP visual encoder, demonstrating the efficacy of our vision-centric supervision tailored for visual outputs. The code will be made publicly available upon acceptance.

In this work, we present a novel direction to build an image tokenizer directly on top of a frozen vision foundation model, which is a largely underexplored area. Specifically, we employ a frozen vision foundation model as the encoder of our tokenizer. To enhance its effectiveness, we introduce two key components: (1) a region-adaptive quantization framework that reduces redundancy in the pre-trained features on regular 2D grids, and (2) a semantic reconstruction objective that aligns the tokenizer’s outputs with the foundation model’s representations to preserve semantic fidelity. Based on these designs, our proposed image tokenizer, \textbf{\ours}, achieves substantial improvements in image reconstruction and generation quality, while also enhancing token efficiency. It further boosts autoregressive (AR) generation---achieving a gFID of \textbf{1.36} on ImageNet benchmarks, while accelerating model convergence by \textbf{three times}, and enabling high-fidelity class-conditional synthesis without the need for classifier-free guidance (CFG). The code is available at \href{https://github.com/CVMI-Lab/VFMTok}{https://github.com/CVMI-Lab/VFMTok}.

In this paper, we propose a new query-based detection framework for crowd detection. Previous query-based detectors suffer from two drawbacks: first, multiple predictions will be inferred for a single object, typically in crowded scenes; second, the performance saturates as the depth of the decoding stage increases. Benefiting from the nature of the one-to-one label assignment rule, we propose a progressive predicting method to address the above issues. Specifically, we first select accepted queries prone to generate true positive predictions, then refine the rest noisy queries according to the previously accepted predictions. Experiments show that our method can significantly boost the performance of query-based detectors in crowded scenes. Equipped with our approach, Sparse RCNN achieves 92.0% AP, 41.4% MR−2and 83.2% JI on the challenging CrowdHuman [35] dataset, outperforming the box-based method MIP [8] that specifies in handling crowded scenarios. Moreover, the proposed method, robust to crowdedness, can still obtain consistent improvements on moderately and slightly crowded datasets like CityPersons [47] and COCO [26]. Code will be made publicly available at https://github.com/megvii-model/Iter-E2EDET.

In this paper, we tackle the problem of learning visual representations from unlabeled scene-centric data. Existing works have demonstrated the potential of utilizing the underlying complex structure within scene-centric data; still, they commonly rely on hand-crafted objectness priors or specialized pretext tasks to build a learning framework, which may harm generalizability. Instead, we propose contrastive learning from data-driven semantic slots, namely SlotCon, for joint semantic grouping and representation learning. The semantic grouping is performed by assigning pixels to a set of learnable prototypes, which can adapt to each sample by attentive pooling over the feature and form new slots. Based on the learned data-dependent slots, a contrastive objective is employed for representation learning, which enhances the discriminability of features, and conversely facilitates grouping semantically coherent pixels together. Compared with previous efforts, by simultaneously optimizing the two coupled objectives of semantic grouping and contrastive learning, our approach bypasses the disadvantages of hand-crafted priors and is able to learn object/group-level representations from scene-centric images. Experiments show our approach effectively decomposes complex scenes into semantic groups for feature learning and significantly benefits downstream tasks, including object detection, instance segmentation, and semantic segmentation. Code is available at: https://github.com/CVMI-Lab/SlotCon.

We propose a simple yet effective proposal-based object detector, aiming at detecting highly-overlapped instances in crowded scenes. The key of our approach is to let each proposal predict a set of correlated instances rather than a single one in previous proposal-based frameworks. Equipped with new techniques such as EMD Loss and Set NMS, our detector can effectively handle the difficulty of detecting highly overlapped objects. On a FPN-Res50 baseline, our detector can obtain 4.9\% AP gains on challenging CrowdHuman dataset and 1.0\% $\text{MR}^{-2}$ improvements on CityPersons dataset, without bells and whistles. Moreover, on less crowed datasets like COCO, our approach can still achieve moderate improvement, suggesting the proposed method is robust to crowdedness.

Finding what is and what is not a salient object can be helpful in developing better features and models in salient object detection (SOD). In this paper, we investigate the images that are selected and discarded in constructing a new SOD dataset and find that many similar candidates, complex shape and low objectness are three main attributes of many non-salient objects. Moreover, objects may have diversified attributes that make them salient. As a result, we propose a novel salient object detector by ensembling linear exemplar regressors. We first select reliable foreground and background seeds using the boundary prior and then adopt locally linear embedding (LLE) to conduct manifold-preserving foregroundness propagation. In this manner, a foregroundness map can be generated to roughly pop-out salient objects and suppress non-salient ones with many similar candidates. Moreover, we extract the shape, foregroundness and attention descriptors to characterize the extracted object proposals, and a linear exemplar regressor is trained to encode how to detect salient proposals in a specific image. Finally, various linear exemplar regressors are ensembled to form a single detector that adapts to various scenarios. Extensive experimental results on 5 dataset and the new SOD dataset show that our approach outperforms 9 state-of-art methods.

This paper proposes a novel approach for segmenting primary video objects by using Complementary Convolutional Neural Networks (CCNN) and neighborhood reversible flow. The proposed approach first pre-trains CCNN on massive images with manually annotated salient objects in an end-to-end manner, and the trained CCNN has two separate branches that simultaneously handle two complementary tasks, i.e., foregroundness and backgroundness estimation. By applying CCNN on each video frame, the spatial foregroundness and backgroundness maps can be initialized, which are then propagated between various frames so as to segment primary video objects and suppress distractors. To enforce efficient temporal propagation, we divide each frame into superpixels and construct neighborhood reversible flow that reflects the most reliable temporal correspondences between superpixels in far-away frames. Within such flow, the initialized foregroundness and backgroundness can be efficiently and accurately propagated along the temporal axis so that primary video objects gradually pop-out and distractors are well suppressed. Extensive experimental results on three video datasets show that the proposed approach achieves impressive performance in comparisons with 18 state-of-the-art models.

Recently, CNN-based models have achieved remarkable success in image-based salient object detection (SOD). In these models, a key issue is to find a proper network architecture that best fits for the task of SOD. Toward this end, this paper proposes two-stream fixation-semantic CNNs, whose architecture is inspired by the fact that salient objects in complex images can be unambiguously annotated by selecting the pre-segmented semantic objects that receive the highest fixation density in eye-tracking experiments. In the two-stream CNNs, a fixation stream is pre-trained on eye-tracking data whose architecture well fits for the task of fixation prediction, and a semantic stream is pre-trained on images with semantic tags that has a proper architecture for semantic perception. By fusing these two streams into an inception-segmentation module and jointly fine-tuning them on images with manually annotated salient objects, the proposed networks show impressive performance in segmenting salient objects. Experimental results show that our approach outperforms 10 state-of-the-art models (5 deep, 5 non-deep) on 4 datasets.