Computer Vision

MagicDriveDiT generates high-resolution, long street-view videos with precise control, exceeding limitations of previous methods in autonomous driving.

VBench++: A new benchmark suite meticulously evaluates video generative models across 16 diverse dimensions, aligning with human perception for improved model development and fairer comparisons.

StyleCodes enables easy style sharing for image generation by encoding styles as compact strings, enhancing control and collaboration while minimizing quality loss.

SAMURAI enhances the Segment Anything Model 2 for real-time, zero-shot visual object tracking by incorporating motion-aware memory and motion modeling, significantly improving accuracy and robustness.

ITACLIP boosts training-free semantic segmentation by architecturally enhancing CLIP, integrating LLM-generated class descriptions, and employing image engineering; achieving state-of-the-art results.

Generative World Explorer (Genex) enables agents to imaginatively explore environments, updating beliefs with generated observations for better decision-making.

Researchers have developed Continuous Speculative Decoding, boosting autoregressive image generation speed by up to 2.33x while maintaining image quality.

AnimateAnything: A unified approach enabling precise & consistent video manipulation via a novel optical flow representation and frequency stabilization.

SEAGULL: A novel network uses vision-language instruction tuning to assess image quality for regions of interest (ROIs) with high accuracy, leveraging masks and a new dataset for fine-grained IQA.

Boosting video temporal grounding, NumPro empowers Vid-LLMs by adding frame numbers, making temporal localization as easy as flipping through manga.

FitDiT boosts virtual try-on realism by enhancing garment details via Diffusion Transformers, improving texture and size accuracy for high-fidelity virtual fashion.

MagicQuill: an intelligent interactive image editing system enabling intuitive, precise image edits via brushstrokes and real-time intent prediction by a multimodal LLM.

Sharingan extracts user action sequences from desktop recordings using novel VLM-based methods, achieving 70-80% accuracy and enabling RPA.

EgoVid-5M: First high-quality dataset for egocentric video generation, enabling realistic human-centric world simulations.

WaLa: a billion-parameter 3D generative model using wavelet encodings achieves state-of-the-art results, generating high-quality 3D shapes in seconds.

GaussianAnything: Interactive point cloud latent diffusion enables high-quality, editable 3D models from images or text, overcoming existing 3D generation limitations.

SAMPart3D: Zero-shot 3D part segmentation across granularities, scaling to large datasets & handling part ambiguity.

OmniEdit, a novel instruction-based image editing model, surpasses existing methods by leveraging specialist supervision and high-quality data, achieving superior performance across diverse editing ta…

Edify Image: groundbreaking pixel-perfect photorealistic image generation using cascaded pixel-space diffusion models with a novel Laplacian diffusion process, enabling diverse applications including …

Add-it: Training-free object insertion in images using pretrained diffusion models by cleverly balancing information from the scene, text prompt, and generated image, achieving state-of-the-art result…