Trending Research

Big Transfer (BiT): General Visual Representation Learning

24 Dec 2019google-research/big_transfer

We conduct detailed analysis of the main components that lead to high transfer performance.

 SOTA for Image Classification on ObjectNet (Bounding Box) (using extra training data)

FEW-SHOT LEARNING FINE-GRAINED IMAGE CLASSIFICATION REPRESENTATION LEARNING

495
2.40 stars / hour

Point2Mesh: A Self-Prior for Deformable Meshes

22 May 2020ranahanocka/point2mesh

We optimize the network weights to deform an initial mesh to shrink-wrap a single input point cloud.

108
1.89 stars / hour

Instance-aware Image Colorization

21 May 2020ericsujw/InstColorization

Previous methods leverage the deep neural network to map input grayscale images to plausible color outputs directly.

COLORIZATION

137
0.65 stars / hour

BERTweet: A pre-trained language model for English Tweets

20 May 2020VinAIResearch/BERTweet

We present BERTweet, the first public large-scale pre-trained language model for English Tweets.

LANGUAGE MODELLING NAMED ENTITY RECOGNITION PART-OF-SPEECH TAGGING TEXT CLASSIFICATION

90
0.27 stars / hour

Encoder-Decoder with Atrous Separable Convolution for Semantic Image Segmentation

ECCV 2018 ayoolaolafenwa/PixelLib

The former networks are able to encode multi-scale contextual information by probing the incoming features with filters or pooling operations at multiple rates and multiple effective fields-of-view, while the latter networks can capture sharper object boundaries by gradually recovering the spatial information.

IMAGE CLASSIFICATION LESION SEGMENTATION SEMANTIC SEGMENTATION

89
0.49 stars / hour

ResNeSt: Split-Attention Networks

19 Apr 2020zhanghang1989/ResNeSt

While image classification models have recently continued to advance, most downstream applications such as object detection and semantic segmentation still employ ResNet variants as the backbone network due to their simple and modular structure.

IMAGE CLASSIFICATION INSTANCE SEGMENTATION OBJECT DETECTION PANOPTIC SEGMENTATION

1,635
0.45 stars / hour

Background Matting: The World is Your Green Screen

1 Apr 2020senguptaumd/Background-Matting

To bridge the domain gap to real imagery with no labeling, we train another matting network guided by the first network and by a discriminator that judges the quality of composites.

IMAGE MATTING

3,028
0.45 stars / hour

A Baseline for 3D Multi-Object Tracking

9 Jul 2019xinshuoweng/AB3DMOT

Although our baseline system is a straightforward combination of standard methods, we obtain the state-of-the-art results.

3D MULTI-OBJECT TRACKING AUTONOMOUS DRIVING

753
0.34 stars / hour