Transformers have catalyzed advancements in computer vision and natural language processing (NLP) fields. However, substantial computational complexity poses limitations for their application in long-context tasks, such as high-resolution image generation. This paper introduces a series of architectures adapted from the RWKV model used in the NLP, with requisite modifications tailored for diffusion model applied to image generation tasks, referred to as Diffusion-RWKV. Similar to the diffusion with Transformers, our model is designed to efficiently handle patchnified inputs in a sequence with extra conditions, while also scaling up effectively, accommodating both large-scale parameters and extensive datasets. Its distinctive advantage manifests in its reduced spatial aggregation complexity, rendering it exceptionally adept at processing high-resolution images, thereby eliminating the necessity for windowing or group cached operations. Experimental results on both condition and unconditional image generation tasks demonstrate that Diffison-RWKV achieves performance on par with or surpasses existing CNN or Transformer-based diffusion models in FID and IS metrics while significantly reducing total computation FLOP usage.
This paper presents a new exploration into a category of diffusion models built upon state space architecture. We endeavor to train diffusion models for image data, wherein the traditional U-Net backbone is supplanted by a state space backbone, functioning on raw patches or latent space. Given its notable efficacy in accommodating long-range dependencies, Diffusion State Space Models (DiS) are distinguished by treating all inputs including time, condition, and noisy image patches as tokens. Our assessment of DiS encompasses both unconditional and class-conditional image generation scenarios, revealing that DiS exhibits comparable, if not superior, performance to CNN-based or Transformer-based U-Net architectures of commensurate size. Furthermore, we analyze the scalability of DiS, gauged by the forward pass complexity quantified in Gflops. DiS models with higher Gflops, achieved through augmentation of depth/width or augmentation of input tokens, consistently demonstrate lower FID. In addition to demonstrating commendable scalability characteristics, DiS-H/2 models in latent space achieve performance levels akin to prior diffusion models on class-conditional ImageNet benchmarks at the resolution of 256$\times$256 and 512$\times$512, while significantly reducing the computational burden. The code and models are available at: https://github.com/feizc/DiS.
Recent real-time semantic segmentation methods usually adopt an additional semantic branch to pursue rich long-range context. However, the additional branch incurs undesirable computational overhead and slows inference speed. To eliminate this dilemma, we propose SCTNet, a single branch CNN with transformer semantic information for real-time segmentation. SCTNet enjoys the rich semantic representations of an inference-free semantic branch while retaining the high efficiency of lightweight single branch CNN. SCTNet utilizes a transformer as the training-only semantic branch considering its superb ability to extract long-range context. With the help of the proposed transformer-like CNN block CFBlock and the semantic information alignment module, SCTNet could capture the rich semantic information from the transformer branch in training. During the inference, only the single branch CNN needs to be deployed. We conduct extensive experiments on Cityscapes, ADE20K, and COCO-Stuff-10K, and the results show that our method achieves the new state-of-the-art performance. The code and model is available at https://github.com/xzz777/SCTNet
Pre-training has emerged as an effective technique for learning powerful person representations. Most existing methods have shown that pre-training on pure-vision large-scale datasets like ImageNet and LUPerson has achieved remarkable performance. However, solely relying on visual information, the absence of robust explicit indicators poses a challenge for these methods to learn discriminative person representations. Drawing inspiration from the intrinsic fine-grained attribute indicators of person descriptions, we explore introducing the language modality into person representation learning. To this end, we propose a novel language-image pre-training framework for person representation learning, termed PLIP. To explicitly build fine-grained cross-modal associations, we specifically design three pretext tasks, \ie semantic-fused image colorization, visual-fused attributes prediction, and vision-language matching. In addition, due to the lack of an appropriate dataset, we present a large-scale person dataset named SYNTH-PEDES, where the Stylish Pedestrian Attributes-union Captioning method is proposed to synthesize diverse textual descriptions. We pre-train PLIP on SYNTH-PEDES and evaluate our model by spanning downstream tasks such as text-based Re-ID, image-based Re-ID, and person attribute recognition. Extensive experiments demonstrate that our model not only significantly improves existing methods on all these tasks, but also shows great ability in the few-shot and domain generalization settings. The code, dataset and weights will be released at~\url{https://github.com/Zplusdragon/PLIP}
Since the fully convolutional network has achieved great success in semantic segmentation, lots of works have been proposed focusing on extracting discriminative pixel feature representations. However, we observe that existing methods still suffer from two typical challenges, i.e. (i) large intra-class feature variation in different scenes, (ii) small inter-class feature distinction in the same scene. In this paper, we first rethink semantic segmentation from a perspective of similarity between pixels and class centers. Each weight vector of the segmentation head represents its corresponding semantic class in the whole dataset, which can be regarded as the embedding of the class center. Thus, the pixel-wise classification amounts to computing similarity in the final feature space between pixels and the class centers. Under this novel view, we propose a Class Center Similarity layer (CCS layer) to address the above-mentioned challenges by generating adaptive class centers conditioned on different scenes and supervising the similarities between class centers. It utilizes a Adaptive Class Center Module (ACCM) to generate class centers conditioned on each scene, which adapt the large intra-class variation between different scenes. Specially designed loss functions are introduced to control both inter-class and intra-class distances based on predicted center-to-center and pixel-to-center similarity, respectively. Finally, the CCS layer outputs the processed pixel-to-center similarity as the segmentation prediction. Extensive experiments demonstrate that our model performs favourably against the state-of-the-art CNN-based methods.
Predicting the future motion of road participants is crucial for autonomous driving but is extremely challenging due to staggering motion uncertainty. Recently, most motion forecasting methods resort to the goal-based strategy, i.e., predicting endpoints of motion trajectories as conditions to regress the entire trajectories, so that the search space of solution can be reduced. However, accurate goal coordinates are hard to predict and evaluate. In addition, the point representation of the destination limits the utilization of a rich road context, leading to inaccurate prediction results in many cases. Goal area, i.e., the possible destination area, rather than goal coordinate, could provide a more soft constraint for searching potential trajectories by involving more tolerance and guidance. In view of this, we propose a new goal area-based framework, named Goal Area Network (GANet), for motion forecasting, which models goal areas rather than exact goal coordinates as preconditions for trajectory prediction, performing more robustly and accurately. Specifically, we propose a GoICrop (Goal Area of Interest) operator to effectively extract semantic lane features in goal areas and model actors' future interactions, which benefits a lot for future trajectory estimations. GANet ranks the 1st on the leaderboard of Argoverse Challenge among all public literature (till the paper submission), and its source codes will be released.
Face manipulation techniques develop rapidly and arouse widespread public concerns. Despite that vanilla convolutional neural networks achieve acceptable performance, they suffer from the overfitting issue. To relieve this issue, there is a trend to introduce some erasing-based augmentations. We find that these methods indeed attempt to implicitly induce more consistent representations for different augmentations via assigning the same label for different augmented images. However, due to the lack of explicit regularization, the consistency between different representations is less satisfactory. Therefore, we constrain the consistency of different representations explicitly and propose a simple yet effective framework, COnsistent REpresentation Learning (CORE). Specifically, we first capture the different representations with different augmentations, then regularize the cosine distance of the representations to enhance the consistency. Extensive experiments (in-dataset and cross-dataset) demonstrate that CORE performs favorably against state-of-the-art face forgery detection methods.
Most existing real-time deep models trained with each frame independently may produce inconsistent results across the temporal axis when tested on a video sequence. A few methods take the correlations in the video sequence into account,e.g., by propagating the results to the neighboring frames using optical flow or extracting frame representations using multi-frame information, which may lead to inaccurate results or unbalanced latency. In this work, we focus on improving the temporal consistency without introducing computation overhead in inference. To this end, we perform inference at each frame. Temporal consistency is achieved by learning from video frames with extra constraints during the training phase. introduced for inference. We propose several techniques to learn from the video sequence, including a temporal consistency loss and online/offline knowledge distillation methods. On the task of semantic video segmentation, weighing among accuracy, temporal smoothness, and efficiency, our proposed method outperforms keyframe-based methods and a few baseline methods that are trained with each frame independently, on datasets including Cityscapes, Camvid, and 300VW-Mask. We further apply our training method to video instance segmentation on YouTubeVISand develop an application of portrait matting in video sequences, by segmenting temporally consistent instance-level trimaps across frames. Experiments show superior qualitative and quantitative results. Code is available at: https://git.io/vidseg.