My Vision, Computer Vision

Talking to DINO: Bridging Self-Supervised Vision Backbones with Language for Open-Vocabulary SegmentationOpen-Vocabulary Segmentation (OVS) aims at segmenting images from free-form textual concepts without predefined training classes. While existing vision-language models such as CLIP can generate segmentation masks by leveraging coarse spatial information frarxiv.orgAuthor : Barsellotti, Luca, ..

LAVT: Language-Aware Vision Transformer for Referring Image SegmentationReferring image segmentation is a fundamental vision-language task that aims to segment out an object referred to by a natural language expression from an image. One of the key challenges behind this task is leveraging the referring expression for highligharxiv.orgAuthor : Yang, Zhao, et alJournal : CVPR 2022Keyword : PWAM, ..

A Survey on Hallucination in Large Vision-Language ModelsRecent development of Large Vision-Language Models (LVLMs) has attracted growing attention within the AI landscape for its practical implementation potential. However, ``hallucination'', or more specifically, the misalignment between factual visual contentarxiv.org Author : Liu, Hanchao, et al.Journal : ArxivKeyword : Survey, Vision Langau..

GRES: Generalized Referring Expression SegmentationReferring Expression Segmentation (RES) aims to generate a segmentation mask for the object described by a given language expression. Existing classic RES datasets and methods commonly support single-target expressions only, i.e., one expression refers toarxiv.orgAuthor : Liu, Chang, Henghui Ding, and Xudong Jiang.Journal : CVPR 2023Keyword : Re..

DINOv2: Learning Robust Visual Features without SupervisionThe recent breakthroughs in natural language processing for model pretraining on large quantities of data have opened the way for similar foundation models in computer vision. These models could greatly simplify the use of images in any system by producingarxiv.org Author : MLAOquab, Maxime, et al.Journal : ArxivKeyword : dinov2Published..

MoVE-KD: Knowledge Distillation for VLMs with Mixture of Visual EncodersVisual encoders are fundamental components in vision-language models (VLMs), each showcasing unique strengths derived from various pre-trained visual foundation models. To leverage the various capabilities of these encoders, recent studies incorporate multarxiv.orgAuthor : Cao, Jiajun, et al.Journal : ArxivKeyword : Knowledg..