VIMI: Grounding Video Generation through Multi-modal Instruction

Existing text-to-video diffusion models rely solely on text-only encoders for their pretraining. This limitation stems from the absence of large-scale multimodal prompt video datasets, resulting in a lack of visual grounding and restricting their versatility and application in multimodal integration. To address this, we construct a large-scale multimodal prompt dataset by employing retrieval methods to pair in-context examples with the given text prompts and then utilize a two-stage training strategy to enable diverse video generation tasks within the same model. In the first stage, we propose a multimodal conditional video generation framework for pretraining on these augmented datasets, establishing a foundational model for grounded video generation. Secondly, we finetune the model from the first stage on three video generation tasks, incorporating multi-modal instructions. This process further refines the model's ability to handle diverse inputs and tasks, ensuring seamless integration of multi-modal information. After this two-stage train-ing process, VIMI demonstrates multimodal understanding capabilities, producing contextually rich and personalized videos grounded in the provided inputs, as shown in Figure 1. Compared to previous visual grounded video generation methods, VIMI can synthesize consistent and temporally coherent videos with large motion while retaining the semantic control. Lastly, VIMI also achieves state-of-the-art text-to-video generation results on UCF101 benchmark.

Panda-70M: Captioning 70M Videos with Multiple Cross-Modality Teachers

The quality of the data and annotation upper-bounds the quality of a downstream model. While there exist large text corpora and image-text pairs, high-quality video-text data is much harder to collect. First of all, manual labeling is more time-consuming, as it requires an annotator to watch an entire video. Second, videos have a temporal dimension, consisting of several scenes stacked together, and showing multiple actions. Accordingly, to establish a video dataset with high-quality captions, we propose an automatic approach leveraging multimodal inputs, such as textual video description, subtitles, and individual video frames. Specifically, we curate 3.8M high-resolution videos from the publicly available HD-VILA-100M dataset. We then split them into semantically consistent video clips, and apply multiple cross-modality teacher models to obtain captions for each video. Next, we finetune a retrieval model on a small subset where the best caption of each video is manually selected and then employ the model in the whole dataset to select the best caption as the annotation. In this way, we get 70M videos paired with high-quality text captions. We dub the dataset as Panda-70M. We show the value of the proposed dataset on three downstream tasks: video captioning, video and text retrieval, and text-driven video generation. The models trained on the proposed data score substantially better on the majority of metrics across all the tasks.

Snap Video: Scaled Spatiotemporal Transformers for Text-to-Video Synthesis

Contemporary models for generating images show remarkable quality and versatility. Swayed by these advantages, the research community repurposes them to generate videos. Since video content is highly redundant, we argue that naively bringing advances of image models to the video generation domain reduces motion fidelity, visual quality and impairs scalability. In this work, we build Snap Video, a video-first model that systematically addresses these challenges. To do that, we first extend the EDM framework to take into account spatially and temporally redundant pixels and naturally support video generation. Second, we show that a U-Net - a workhorse behind image generation - scales poorly when generating videos, requiring significant computational overhead. Hence, we propose a new transformer-based architecture that trains 3.31 times faster than U-Nets (and is ~4.5 faster at inference). This allows us to efficiently train a text-to-video model with billions of parameters for the first time, reach state-of-the-art results on a number of benchmarks, and generate videos with substantially higher quality, temporal consistency, and motion complexity. The user studies showed that our model was favored by a large margin over the most recent methods. See our website at https://snap-research.github.io/snapvideo/.

Motion-Conditioned Diffusion Model for Controllable Video Synthesis

Recent advancements in diffusion models have greatly improved the quality and diversity of synthesized content. To harness the expressive power of diffusion models, researchers have explored various controllable mechanisms that allow users to intuitively guide the content synthesis process. Although the latest efforts have primarily focused on video synthesis, there has been a lack of effective methods for controlling and describing desired content and motion. In response to this gap, we introduce MCDiff, a conditional diffusion model that generates a video from a starting image frame and a set of strokes, which allow users to specify the intended content and dynamics for synthesis. To tackle the ambiguity of sparse motion inputs and achieve better synthesis quality, MCDiff first utilizes a flow completion model to predict the dense video motion based on the semantic understanding of the video frame and the sparse motion control. Then, the diffusion model synthesizes high-quality future frames to form the output video. We qualitatively and quantitatively show that MCDiff achieves the state-the-of-art visual quality in stroke-guided controllable video synthesis. Additional experiments on MPII Human Pose further exhibit the capability of our model on diverse content and motion synthesis.

Hard Samples Rectification for Unsupervised Cross-domain Person Re-identification

Person re-identification (re-ID) has received great success with the supervised learning methods. However, the task of unsupervised cross-domain re-ID is still challenging. In this paper, we propose a Hard Samples Rectification (HSR) learning scheme which resolves the weakness of original clustering-based methods being vulnerable to the hard positive and negative samples in the target unlabelled dataset. Our HSR contains two parts, an inter-camera mining method that helps recognize a person under different views (hard positive) and a part-based homogeneity technique that makes the model discriminate different persons but with similar appearance (hard negative). By rectifying those two hard cases, the re-ID model can learn effectively and achieve promising results on two large-scale benchmarks.

Incremental False Negative Detection for Contrastive Learning

Self-supervised learning has recently shown great potential in vision tasks via contrastive learning, which aims to discriminate each image, or instance, in the dataset. However, such instance-level learning ignores the semantic relationship between instances and repels the anchor equally from the semantically similar samples, termed as false negatives. In this work, we first empirically highlight that the unfavorable effect from false negatives is more significant for the datasets containing images with more semantic concepts. To address the issue, we introduce a novel incremental false negative detection for self-supervised contrastive learning. Following the training process, when the encoder is gradually better-trained and the embedding space becomes more semantically structural, our method incrementally detects more reliable false negatives. Subsequently, during contrastive learning, we discuss two strategies to explicitly remove the detected false negatives. Extensive experiments show that our proposed method outperforms other self-supervised contrastive learning frameworks on multiple benchmarks within a limited compute.

Viewpoint-Aware Channel-Wise Attentive Network for Vehicle Re-Identification

Vehicle re-identification (re-ID) matches images of the same vehicle across different cameras. It is fundamentally challenging because the dramatically different appearance caused by different viewpoints would make the framework fail to match two vehicles of the same identity. Most existing works solved the problem by extracting viewpoint-aware feature via spatial attention mechanism, which, yet, usually suffers from noisy generated attention map or otherwise requires expensive keypoint labels to improve the quality. In this work, we propose Viewpoint-aware Channel-wise Attention Mechanism (VCAM) by observing the attention mechanism from a different aspect. Our VCAM enables the feature learning framework channel-wisely reweighing the importance of each feature maps according to the "viewpoint" of input vehicle. Extensive experiments validate the effectiveness of the proposed method and show that we perform favorably against state-of-the-arts methods on the public VeRi-776 dataset and obtain promising results on the 2020 AI City Challenge. We also conduct other experiments to demonstrate the interpretability of how our VCAM practically assists the learning framework.

Orientation-aware Vehicle Re-identification with Semantics-guided Part Attention Network

Vehicle re-identification (re-ID) focuses on matching images of the same vehicle across different cameras. It is fundamentally challenging because differences between vehicles are sometimes subtle. While several studies incorporate spatial-attention mechanisms to help vehicle re-ID, they often require expensive keypoint labels or suffer from noisy attention mask if not trained with expensive labels. In this work, we propose a dedicated Semantics-guided Part Attention Network (SPAN) to robustly predict part attention masks for different views of vehicles given only image-level semantic labels during training. With the help of part attention masks, we can extract discriminative features in each part separately. Then we introduce Co-occurrence Part-attentive Distance Metric (CPDM) which places greater emphasis on co-occurrence vehicle parts when evaluating the feature distance of two images. Extensive experiments validate the effectiveness of the proposed method and show that our framework outperforms the state-of-the-art approaches.