딥러닝 공부방

DETIC, Detecting Twenty thousand Classes using Image-label Supervision https://arxiv.org/abs/2201.02605 Detecting Twenty-thousand Classes using Image-level Supervision Current object detectors are limited in vocabulary size due to the small scale of detection datasets. Image classifiers, on the other hand, reason about much larger vocabularies, as their datasets are larger and easier to collect...

Open-Vocabulary Object Detection via Vision and Language Knowledge Distillation, ViLD https://arxiv.org/abs/2104.13921 Open-vocabulary Object Detection via Vision and Language Knowledge Distillation We aim at advancing open-vocabulary object detection, which detects objects described by arbitrary text inputs. The fundamental challenge is the availability of training data. Existing object detecti..

https://arxiv.org/abs/2112.01071 DenseCLIP: Extract Free Dense Labels from CLIP Contrastive Language-Image Pre-training (CLIP) has made a remarkable breakthrough in open-vocabulary zero-shot image recognition. Many recent studies leverage the pre-trained CLIP models for image-level classification and manipulation. In this paper, we fu arxiv.org CLIP을 segmentation에 적용한 논문. clip이 학습한 정보를 segmentat..

Generalized Category Discovery https://arxiv.org/abs/2201.02609 Generalized Category Discovery In this paper, we consider a highly general image recognition setting wherein, given a labelled and unlabelled set of images, the task is to categorize all images in the unlabelled set. Here, the unlabelled images may come from labelled classes or from nov arxiv.org 새로운 task를 제안한다. training set에 포함되어 있..

f-VAEGAN-D2: A Feature Generating Framework for Any-Shot Learning https://arxiv.org/abs/1903.10132 f-VAEGAN-D2: A Feature Generating Framework for Any-Shot Learning When labeled training data is scarce, a promising data augmentation approach is to generate visual features of unknown classes using their attributes. To learn the class conditional distribution of CNN features, these models rely on ..

Decoupling Zero-Shot Semantic Segmentation https://arxiv.org/abs/2112.07910 Decoupling Zero-Shot Semantic Segmentation Zero-shot semantic segmentation (ZS3) aims to segment the novel categories that have not been seen in the training. Existing works formulate ZS3 as a pixel-level zero-shot classification problem, and transfer semantic knowledge from seen classes to unseen arxiv.org MaskFormer와 C..

A Simple Baseline for Zero-Shot Semantic Segmentation with Pre-trained Vision-language Model https://arxiv.org/pdf/2112.14757.pdf CLIP을 zero-shot semantic segmentation에 적용한 논문. MaskFormer로 binary mask를 생성하고 생성한 mask에 대해 mask classification으로 prediction을 수행한다. classifier의 weight를 CLIP의 pre-trained text representation로 사용. 따라서 unseen으로 zero-shot이 가능하다.

https://arxiv.org/abs/2107.13807 FREE: Feature Refinement for Generalized Zero-Shot Learning Generalized zero-shot learning (GZSL) has achieved significant progress, with many efforts dedicated to overcoming the problems of visual-semantic domain gap and seen-unseen bias. However, most existing methods directly use feature extraction models traine arxiv.org FREE, Feature Refinement for Generaliz..

Task-Indenpendent Knowledge Makes for Transferable Represenations for Generalized Zero-Shot Learning https://arxiv.org/abs/2104.01832 Task-Independent Knowledge Makes for Transferable Representations for Generalized Zero-Shot Learning Generalized Zero-Shot Learning (GZSL) targets recognizing new categories by learning transferable image representations. Existing methods find that, by aligning im..

Zero-Shot Learning via Contrastive Learning on Dual Knowledge Graphs https://ieeexplore.ieee.org/document/9607851 Zero-Shot Learning via Contrastive Learning on Dual Knowledge Graphs Graph Convolutional Networks (GCNs), which can integrate both explicit knowledge and implicit knowledge together, have shown effectively for zero-shot learning problems. Previous GCN-based methods generally leverage..