Comments (2)
Problem solved.
A few tips for evaluating MOT results using this benchmark:
The correct file organization should be: ~/AB3DMOT/results/{result_folder}/data/{seq_name}.txt, where {result_folder} is the parameter in the command.
For example when running python evaluation/evaluate_kitti3dmot.py car_3d_det_val
, {result_folder} is car_3d_det_val
. And {seq_name} is simply name of the sequences, for example '0000.txt'. The key thing is that the folder to look for the result is hardcoded in the code at here, where variable t_sha
is the same as {result_folder} mentioned above.
Another problem is that by default, you must evaluate all sequences at the same time, otherwise, it will return a 'problem with data format' indication.
If you only want to evaluate part of the sequences, you can hardcode the sequence you want before this line. For example, if you only want to evaluate sequence 17, you should add self.sequence_name = ['0017']
before the line indicated above.
@xinshuoweng I suggest that you make some modifications or at least clarify this in README coz it's quite unreasonable. For example, there is no 'Car' category in sequence 17 thus cannot be evaluated. What's more, some learning-based MOT methods will split the training set further into sequences for training and evaluation thus will not evaluate all sequences at once.
from ab3dmot.
Correct findings! Although I have already pointed out how to run the code in the readme files (e.g., python main.py car_3d_det_val).
For evaluating different sequences, as you have already found out, it is very simple to change and modify for different sequences. In my cases, I have to evaluate all the sequences and is thus reasonable to have the provided evaluation script and that is also the default for the KITTI dataset. If you are not doing default, you have to modify the code.
from ab3dmot.
Related Issues (20)
- quick start no datasets... HOT 9
- ValueError: not enough values to unpack (expected 2, got 1) HOT 10
- Ego Motion
- Evaluation on Custom Tracking HOT 2
- Is your kalman gain updated with status, or is it a fixed value? Thanks!
- Is there some tips about how to write one script to feed my Robosense point cloud?
- HOTA
- C++ implement HOT 1
- Converting nuScenes detections to KITTI format missing correspondence files HOT 3
- Can you share the new change lists of the code this year?
- Question about the KITTI detections. HOT 2
- Project dependencies may have API risk issues
- Question about visualization HOT 2
- carefully followed provided installation instructions, got FileNotFoundError HOT 7
- real-time data HOT 4
- 无法评估的问题
- Detector/Training procedure to get the provided detections HOT 1
- 请问是否修改过工程 HOT 1
- A question about detection results which in data/KITTI/detection/
- Speed optimization and parameter choice for custom detector
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 ab3dmot.