Concerns with enabling computers to interpret and understand visual inputs, such as images and videos.
The article introduces MatSynth, a dataset comprised of over 4,000 ultra-high resolution PBR materials. These materials are key for defining the interaction of light on the surfaces of virtual objects….
Continue readingThe LEGO model is a multi-modal model that emphasizes both global and local information across different modalities. Unlike existing models, which focus mainly on global information, the LEGO model can…
Continue readingThe article discusses a new concept called ‘prompt-aligned personalization’ for improving the performance of text-to-image models. It addresses the limitations of existing personalization methods, such as compromising the ability to…
Continue readingThis research presents a new shadow model named Gaussian Shadow Casting (GSC). The model can reconstruct 3D neural characters from videos, with improved shadows and shading, even in challenging outdoor…
Continue readingThe article presents a new method for visual dubbing, which is the process of generating lip motions of an actor in a video to synchronize with given audio. The method,…
Continue readingThe article introduces E2GAN, a novel approach for efficient training of GANs for image-to-image translation. This technique uses data distillation from large-scale text-to-image diffusion models, such as Stable Diffusion, for…
Continue readingThe research aims to replicate the success of image-text data for video-language models. The researchers fine-tuned a video-language model from a strong image-language baseline with synthesized instructional data. The adapted…
Continue readingThe article introduces a novel audio-video recognition approach called the Audio-Video Transformer (AVT) that uses effective spatio-temporal representation for improved action recognition. The research reduces cross-modality complexity via an audio-video…
Continue readingThe article introduces ‘FunnyNet-W’, a model that relies on cross- and self-attention for visual, audio, and text data to predict funny moments in videos. Unlike most methods that rely on…
Continue readingThe article introduces ‘Anim-400K’, a large-scale dataset designed to aid in the automated end-to-end dubbing of video content. With 60% of online content published in English and only 18.8% of…
Continue readingLooks like you're using an ad blocker. Please support us by disabling these ad blocker.
