|
- 论文笔记： Pixel-Level Contrastive learning - 知乎
作者结合了pixel-level和instance-level的方法，利用了前者的空间敏感性（spatially sensitive）和后者的分类能力（categorization ability），充分利用了它们各自的优势。
- GitHub - MaverickRen PixelLM: [CVPR 2024] PixelLM is an effective and . . .
We present PixelLM, a novel LMM for pixel-level reasoning and understanding PixelLM proficiently handles tasks with an arbitrary number of open-set targets and diverse reasoning complexities
- Pixel-level and Semantic-level Adjustable Super-resolution: A Dual-LoRA . . .
We present Pixel-level and Semantic-level Adjustable SR (PiSA-SR), which learns two LoRA modules upon the pre-trained stable-diffusion (SD) model to achieve improved and adjustable SR results
- OMG-LLaVA: Bridging Image-level, Object-level, Pixel-level Reasoning . . .
We propose OMG-LLaVA, a new and elegant framework combining powerful pixel-level vision understanding with reasoning abilities It can accept various visual and text prompts for flexible user interaction
- Pixel-level Semantic Correspondence through Layout-aware Representation . . .
It focuses on shared geometries and pixel-level semantics for reliable correspon-dence, capable of handling large-scale object variations and improving robustness across various scenarios
- PixCon: Pixel-Level Contrastive Learning Revisited - MDPI
We prove this by proposing PixCon, a pixel-level contrastive learning framework, and testing different positive matching strategies based on this framework to rediscover the potential of pixel-level learning
- GeoPix: Multi-Modal Large Language Model for Pixel-level Image . . .
In this paper, we propose GeoPix, a RS MLLM that extends image understanding capabilities to the pixel level This is achieved by equipping the MLLM with a mask predictor, which transforms visual features from the vision encoder into masks conditioned on the LLM's segmentation token embeddings
- 【自监督系列】首次探究像素级别的自监督任务 - 知乎
作者认为，pixel-level的方法是对instance-level方法的补充，instance-level方法擅长学习类别信息、整体特征， pixel-level 方法擅长学习spatially sensitive representation，两者结合是最好的，而且计算高效，两个任务是共享一个backbone encoder的。
|
|
|