DenseVLM: A Retrieval and Decoupled Alignment Framework for Open-Vocabulary Dense Prediction

Pre-trained vision-language models (VLMs), such as CLIP, have demonstrated impressive zero-shot recognition capability, but still underperform in dense prediction tasks. Self-distillation recently is emerging as a promising approach for fine-tuning VLMs to better adapt to local regions without requiring extensive annotations. However, previous state-of-the-art approaches often suffer from significant `foreground bias', where models tend to wrongly identify background regions as foreground objects. To alleviate this issue, we propose DenseVLM, a framework designed to learn unbiased region-language alignment from powerful pre-trained VLM representations. By leveraging the pre-trained VLM to retrieve categories for unlabeled regions, DenseVLM effectively decouples the interference between foreground and background region features, ensuring that each region is accurately aligned with its corresponding category. We show that DenseVLM can be seamlessly integrated into open-vocabulary object detection and image segmentation tasks, leading to notable performance improvements. Furthermore, it exhibits promising zero-shot scalability when training on more extensive and diverse datasets.

PR-MIM: Delving Deeper into Partial Reconstruction in Masked Image Modeling

Masked image modeling has achieved great success in learning representations but is limited by the huge computational costs. One cost-saving strategy makes the decoder reconstruct only a subset of masked tokens and throw the others, and we refer to this method as partial reconstruction. However, it also degrades the representation quality. Previous methods mitigate this issue by throwing tokens with minimal information using temporal redundancy inaccessible for static images or attention maps that incur extra costs and complexity. To address these limitations, we propose a progressive reconstruction strategy and a furthest sampling strategy to reconstruct those thrown tokens in an extremely lightweight way instead of completely abandoning them. This approach involves all masked tokens in supervision to ensure adequate pre-training, while maintaining the cost-reduction benefits of partial reconstruction. We validate the effectiveness of the proposed method across various existing frameworks. For example, when throwing 50% patches, we can achieve lossless performance of the ViT-B/16 while saving 28% FLOPs and 36% memory usage compared to standard MAE. Our source code will be made publicly available

Cascade-CLIP: Cascaded Vision-Language Embeddings Alignment for Zero-Shot Semantic Segmentation

Pre-trained vision-language models, e.g., CLIP, have been successfully applied to zero-shot semantic segmentation. Existing CLIP-based approaches primarily utilize visual features from the last layer to align with text embeddings, while they neglect the crucial information in intermediate layers that contain rich object details. However, we find that directly aggregating the multi-level visual features weakens the zero-shot ability for novel classes. The large differences between the visual features from different layers make these features hard to align well with the text embeddings. We resolve this problem by introducing a series of independent decoders to align the multi-level visual features with the text embeddings in a cascaded way, forming a novel but simple framework named Cascade-CLIP. Our Cascade-CLIP is flexible and can be easily applied to existing zero-shot semantic segmentation methods. Experimental results show that our simple Cascade-CLIP achieves superior zero-shot performance on segmentation benchmarks, like COCO-Stuff, Pascal-VOC, and Pascal-Context. Our code is available at: https://github.com/HVision-NKU/Cascade-CLIP

A Decoupled Spatio-Temporal Framework for Skeleton-based Action Segmentation

Effectively modeling discriminative spatio-temporal information is essential for segmenting activities in long action sequences. However, we observe that existing methods are limited in weak spatio-temporal modeling capability due to two forms of decoupled modeling: (i) cascaded interaction couples spatial and temporal modeling, which over-smooths motion modeling over the long sequence, and (ii) joint-shared temporal modeling adopts shared weights to model each joint, ignoring the distinct motion patterns of different joints. We propose a Decoupled Spatio-Temporal Framework (DeST) to address the above issues. Firstly, we decouple the cascaded spatio-temporal interaction to avoid stacking multiple spatio-temporal blocks, while achieving sufficient spatio-temporal interaction. Specifically, DeST performs once unified spatial modeling and divides the spatial features into different groups of subfeatures, which then adaptively interact with temporal features from different layers. Since the different sub-features contain distinct spatial semantics, the model could learn the optimal interaction pattern at each layer. Meanwhile, inspired by the fact that different joints move at different speeds, we propose joint-decoupled temporal modeling, which employs independent trainable weights to capture distinctive temporal features of each joint. On four large-scale benchmarks of different scenes, DeST significantly outperforms current state-of-the-art methods with less computational complexity.

TeMO: Towards Text-Driven 3D Stylization for Multi-Object Meshes

Recent progress in the text-driven 3D stylization of a single object has been considerably promoted by CLIP-based methods. However, the stylization of multi-object 3D scenes is still impeded in that the image-text pairs used for pre-training CLIP mostly consist of an object. Meanwhile, the local details of multiple objects may be susceptible to omission due to the existing supervision manner primarily relying on coarse-grained contrast of image-text pairs. To overcome these challenges, we present a novel framework, dubbed TeMO, to parse multi-object 3D scenes and edit their styles under the contrast supervision at multiple levels. We first propose a Decoupled Graph Attention (DGA) module to distinguishably reinforce the features of 3D surface points. Particularly, a cross-modal graph is constructed to align the object points accurately and noun phrases decoupled from the 3D mesh and textual description. Then, we develop a Cross-Grained Contrast (CGC) supervision system, where a fine-grained loss between the words in the textual description and the randomly rendered images are constructed to complement the coarse-grained loss. Extensive experiments show that our method can synthesize high-quality stylized content and outperform the existing methods over a wide range of multi-object 3D meshes. Our code and results will be made publicly available

End-to-End Streaming Video Temporal Action Segmentation with Reinforce Learning

Temporal Action Segmentation (TAS) from video is a kind of frame recognition task for long video with multiple action classes. As an video understanding task for long videos, current methods typically combine multi-modality action recognition models with temporal models to convert feature sequences to label sequences. This approach can only be applied to offline scenarios, which severely limits the TAS application. Therefore, this paper proposes an end-to-end Streaming Video Temporal Action Segmentation with Reinforce Learning (SVTAS-RL). The end-to-end SVTAS which regard TAS as an action segment clustering task can expand the application scenarios of TAS; and RL is used to alleviate the problem of inconsistent optimization objective and direction. Through extensive experiments, the SVTAS-RL model achieves a competitive performance to the state-of-the-art model of TAS on multiple datasets, and shows greater advantages on the ultra-long video dataset EGTEA. This indicates that our method can replace all current TAS models end-to-end and SVTAS-RL is more suitable for long video TAS. Code is availabel at https://github.com/Thinksky5124/SVTAS.

Streaming Video Temporal Action Segmentation In Real Time

Temporal action segmentation (TAS) is a critical step toward long-term video understanding. Recent studies follow a pattern that builds models based on features instead of raw video picture information. However, we claim those models are trained complicatedly and limit application scenarios. It is hard for them to segment human actions of video in real time because they must work after the full video features are extracted. As the real-time action segmentation task is different from TAS task, we define it as streaming video real-time temporal action segmentation (SVTAS) task. In this paper, we propose a real-time end-to-end multi-modality model for SVTAS task. More specifically, under the circumstances that we cannot get any future information, we segment the current human action of streaming video chunk in real time. Furthermore, the model we propose combines the last steaming video chunk feature extracted by language model with the current image feature extracted by image model to improve the quantity of real-time temporal action segmentation. To the best of our knowledge, it is the first multi-modality real-time temporal action segmentation model. Under the same evaluation criteria as full video temporal action segmentation, our model segments human action in real time with less than 40% of state-of-the-art model computation and achieves 90% of the accuracy of the full video state-of-the-art model.