Comments (2)
Hi, thanks for your attention and insightful question.
- We both choose the BEV feature space to fuse the features of the two modalities, and both choose LSS as the camera feature extractor. MIT-BEVFusion proposes a high-performance multi-task model that achieves remarkable results in 3D detection and MAP semantic segmentation. Our work focuses on the robustness of the multimodal framework, and more work is done on the framework's compatibility with different models and robustness to sensor failures.
- Since in autonomous driving scenarios, most of the objects of interest are on the same level, such as a non-flying car, the importance of z-axis features is relatively low. The BEV feature compresses the z-axis feature to the channel dimension, allowing the model to focus on the perception area of the horizontal plane for more efficient 3D detection.
The above is my personal opinion, welcome to discuss and share your views.
from bevfusion.
Thanks!
from bevfusion.
Related Issues (20)
- Question about nuscenes validation performance HOT 1
- TransFusion-L pretrain model
- only camera test mAP==0? HOT 3
- Can not to reproduce BEVFusion (CenterPoint+VoxelNet+Dual-Swin-T,mAP:64.2,NDS:68.0)
- Memory consumption and inference time
- Question about nuscenes camera2lidar extrinsic
- please install mmdet in mmdetection-2.11.0 HOT 2
- ball_query_ext.so undefined symbol
- Nuscenes detection errors on validation set
- Excuse me.There is a problem with the preprocessing mini's nuscence dataset, how should I solve it? HOT 2
- Camera stream augmentations
- se_block在你们的模块中真的work了吗
- 找不到'work_dirs/mask_rcnn_dbswin-t_fpn_3x_nuim_cocopre/epoch_36.pth' HOT 1
- 能否添加一个可视化的命令
- Can anyone share me with the model in camera stream, which name is 'models/cam_tf_e20.pth'?
- AssertionError: top_pool_forward miss in module _ext while executingpython demo/pcd_demo.py ${PCD_FILE} ${CONFIG_FILE} ${CHECKPOINT_FILE} [--device ${GPU_ID}] [--score-thr ${SCORE_THR}] [--out-dir ${OUT_DIR}]
- Installation problem: mmcv package configuration HOT 4
- DockerFile有问题 HOT 2
- Code does not match the paper
- There's no pretrained model for DAL-tiny HOT 1
Recommend Projects
-
React
A declarative, efficient, and flexible JavaScript library for building user interfaces.
-
Vue.js
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
-
Typescript
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
-
TensorFlow
An Open Source Machine Learning Framework for Everyone
-
Django
The Web framework for perfectionists with deadlines.
-
Laravel
A PHP framework for web artisans
-
D3
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
-
Recommend Topics
-
javascript
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
-
web
Some thing interesting about web. New door for the world.
-
server
A server is a program made to process requests and deliver data to clients.
-
Machine learning
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
-
Visualization
Some thing interesting about visualization, use data art
-
Game
Some thing interesting about game, make everyone happy.
Recommend Org
-
Facebook
We are working to build community through open source technology. NB: members must have two-factor auth.
-
Microsoft
Open source projects and samples from Microsoft.
-
Google
Google ❤️ Open Source for everyone.
-
Alibaba
Alibaba Open Source for everyone
-
D3
Data-Driven Documents codes.
-
Tencent
China tencent open source team.
from bevfusion.