ECCV 2020 Best Paper Award | RAFT: A New Deep Network Architecture For Optical Flow | WITH CODE

Показать описание

This week my interest was directed towards the ECCV2020 that happened last week. Ask any questions or remarks you have in the comments, I will gladly answer to everything!

Subscribe to not miss any AI news and terms clearly vulgarized! Share this to someone who needs to learn more about Artificial Intelligence! Spread knowledge, not germs!

Follow me for more AI content:

The best courses to start and progress in AI:

Join Our Discord channel, Learn AI Together:

Support me on patreon:

Chapters:
0:00 Hey! Tap the Thumbs Up button and Subscribe to help me. You'll learn a lot of cool stuff, I promise.
0:48 Paper explanation
4:18 Examples
4:48 Conclusion

#ComputerVision #ECCV #ECCV2020

Рекомендации по теме

Комментарии

The best channel for AI news. I hope that you can expand the channel even more!

nima

I'm sad that "They even made the code available" is actually something that is worth mentioning these days.

gopietz

I saw this featured by Corridor Digital, and wanted to create a whole video project that exists in the rainbow view of the this technology. Now im finding that this may be outside of my abilities if it involves coding and python. I am primarily an adobe premiere, after effects and photoshop user. Is there no consumer friendly tutorial out there for this technology?

zaneberry

I had a doubt. I tried Farneback OpenCV implementation using GPU (Nvidia gtx 1600) and the results were much faster than the deep learning RAFT implementation. I was confused as while Farneback took only 33 seconds on a 10 second video, RAFT took 95 seconds(RAFT took 0.3 seconds for a single frame) ! Do these deep learning approaches work better in terms of quality only? Why should we use these deep learning approaches?

siftir

So ..now I'm confident enough that some basic coding is enough to get through complex issues.

varaprasad

I wonder how the ground truth for optical flow is obtained? If it is from conventional optical flow model, then how can we say the new model provides more accurate result?

ruanjiayang

What's best one around now? FGVC? This RAFT is bit older.

vodkaru

it's possible to have the code or the github

emmanuelwafo

What does it mean to train on optical flow vs pixels?

aguy

the Article is for 2003 ? it seems old ...

mbseifollahi

ECCV 2020 Best Paper Award | RAFT: A New Deep Network Architecture For Optical Flow | WITH CODE

ECCV 2020 Best Paper Award | RAFT: A New Deep Network Architecture For Optical Flow | WITH CODE

ECCV 2020 Paper Compilation - TUM Visual Computing Lab & Collaborators

ECCV 2022 - CCRL - AIMIA

ECCV 2020 Adversarial Robustness Workshop Keynote

ECCV 2020 Paper NODIS: Neural Ordinary Differential Scene Understanding

ECCV How To: Papers & Presentations

ECCV 2020 60s Quickie: Federated Visual Classification with Real-World Data Distribution

ECCV 2020 Tuesday Welcome

ECCV 2020 SHARP Challenges - Results

[ECCV 2020] A2CL-PT for Weakly-supervised Temporal Activity Localization (short video)

[ECCV 2020 - 1 minute] AttentionNAS: Spatiotemporal Attention Cell Search for Video Classification

(ECCV 2020) Impact of base dataset design on few-shot image classification

[ECCV 2020] Searching Efficient 3D Architectures with Sparse Point-Voxel Convolution

[ECCV 2022] [Best Paper] On the Versatile Uses of Partial Distance Correlation in Deep Learning

EGSR 2020: Closing and Best-Paper Award

InterHand2.6M (3D Interacting Hand Pose Estimation. ECCV 2020)

[ECCV-2020] Simultaneous Detection and Tracking with Motion Modelling for Multiple Object Tracking

ECCV 2020 paper: Learning from Extrinsic and Intrinsic Supervisions for Domain Generalization

Open-Edit (ECCV 2020)

ECCV 2020: 'Shape prior deformation for categorical 6D object pose and size estimation'

Memento: Multimodal Memorability (ECCV 2020 video presentation)

Sea-thru ECCV2020 DEMO

Hidden Footprints (ECCV 2020)

Mask2CAD ECCV 2020 Full Video