-
## 一言でいうと
YOLO新作
single faeture output
mAP/GFLOPSはyolov4より少し弱いくらい ( 細かい条件はそろえていないだろうけど )
yolov4-608 65.7% mAP@0.5 (43.5% AP@0.5:0.95) 128.5BFLOPS
yolov4-512 64.9% mAP@0.5 (43.0% AP@0.5:0.95) 91…
-
## Challenge for RGB Super Resolution (SISR 문제라고 정의함)
- NTIRE 2016-2021 라는 이름으로 6년간 지속해왔음
- 이름: New Trends in Image Restoration and Enhancement workshop
- [CVPR2021 Workshop](https://data.vision.ee…
-
DCPose: Deep Dual Consecutive Network for Human Pose Estimation
Paper: https://arxiv.org/abs/2103.07254
Code: https://github.com/Pose-Group/DCPose
-
Hello, is there a plan to release **fcpose** code as mentioned in the cvpr2021 paper?
many thanks!
-
Hi @bsuleymanov,
The Short-video Face Parsing Challenge Dataset is quite interesting but I can't seem to find any download links on [https://maadaa.ai/cvpr2021-short-video-face-parsing-challenge/ ]…
-
想将CVPR2021的NBNet训练的模型转为.onnx,报错,是否目前仅支持部分网络?
![lALPJxf-v_HmYN3NASvNBU0_1357_299](https://user-images.githubusercontent.com/56590176/156712412-be980126-3971-42d3-9fb4-8dbe29d5df84.png)
Megengine 1…
-
**Is your feature request related to a problem? Please describe.**
[CVPR (Computer Vision and Pattern Recognition)](https://en.wikipedia.org/wiki/Conference_on_Computer_Vision_and_Pattern_Recognition…
-
Hi , I have understand your algorithm. It is very interesting.
However, could you please tell me what is the theoretical support of the decomposed equation of the normal vector N in section 4.2?
And…
-
CVPR2021的分类中,点云分类下:
[15] PAConv: Position Adaptive Convolution with Dynamic Kernel Assembling on Point Clouds(PAConv:点云上具有动态内核组装的位置自适应卷积)
这篇的code链接请改为:
https://github.com/CVMI-Lab/PAConv
Thank…
-
Title: Cross-Modal Center Loss for 3D Cross-Modal Retrieval
Paper Link: https://arxiv.org/abs/2008.03561
Code Link: https://github.com/LongLong-Jing/Cross-Modal-Center-Loss