2
2
[ ![ Pylint] ( https://github.com/TimoK93/ctc-metrics/actions/workflows/pylint.yml/badge.svg )] ( https://github.com/TimoK93/ctc-metrics/actions/workflows/pylint.yml )
3
3
[ ![ Python package] ( https://github.com/TimoK93/ctc-metrics/actions/workflows/python-package.yml/badge.svg )] ( https://github.com/TimoK93/ctc-metrics/actions/workflows/python-package.yml )
4
4
5
- # CTC-Metrics
6
- A python implementation of the metrics used in the
7
- [ Cell-Tracking-Challenge] ( https://celltrackingchallenge.net/ ) . Additionally,
8
- [ CHOTA] ( ... ) ,
9
- and the metrics of the
10
- [ MOTChallenge] ( https://motchallenge.net/ ) are implemented.
5
+ # Py-CTCMetrics
6
+ A python implementation of the metrics used in the paper
7
+ [ CHOTA: A Higher Order Accuracy Metric for Cell Tracking] ( ... ) by
8
+ * Kaiser et al.* . The code is
9
+ designed to evaluate tracking results in the format of the
10
+ [ Cell-Tracking-Challenge] ( https://celltrackingchallenge.net/ ) but can also be used
11
+ for custom results.
12
+ The repository contains the metrics of the
13
+ [ Cell-Tracking-Challenge] ( https://celltrackingchallenge.net/ ) ,
14
+ the [ MOTChallenge] ( https://motchallenge.net/ ) , and the
15
+ [ CHOTA] ( ... ) metric.
16
+
17
+ Detailed descriptions of the metrics can be found in the [ paper] ( ... ) .
11
18
19
+ ---
20
+
21
+ # Features at a Glance
12
22
23
+ - ** Validation** : Check if the tracking results are correctly formatted.
24
+ - ** Evaluation** : Evaluate the tracking results with respect to the ground truth using CTC, MotChallenge, and CHOTA metrics.
25
+ - ** Visualization** : Visualize the tracking results.
26
+ - ** Video Creation** : Create a video of the visualized tracking results for presentation purposes.
27
+ - ** Noise Evaluation** : Evaluate the impact of specific errors on the evaluation metrics.
28
+
13
29
---
14
30
15
- ### Requirements
31
+ ## Requirements
16
32
17
33
We tested the code with ** Python 3.10** . Additional packages that will be
18
34
installed automatically are listed in the [ requirements.txt] ( requirements.txt ) .
19
35
36
+ ---
37
+
20
38
## Installation
21
39
22
40
The package can be installed via pip:
23
41
24
42
``` bash
25
- pip install git+https://github.com/TimoK93/ctc-metrics .git
43
+ pip install git+https://github.com/CellTrackingChallenge/py-ctcmetrics .git
26
44
```
27
45
28
46
or from the source code:
29
47
30
48
``` bash
31
- git clone https://github.com/TimoK93/ctc-metrics
32
- cd ctc-metrics
49
+ git clone https://github.com/CellTrackingChallenge/py-ctcmetrics
50
+ cd py-ctcmetrics
33
51
pip install .
34
52
```
35
53
54
+ ---
55
+
36
56
## Usage
37
57
38
58
@@ -99,7 +119,8 @@ The ```gt``` argument specifies the path to the ground truth directory, and the
99
119
recursively evaluates all sequences in the directory.
100
120
101
121
Per default, the code is executed using multiple processes with one process per
102
- available CPU core. Multiprocessing increases the execution time but also
122
+ available CPU core. Multiprocessing decreases the execution time but also
123
+ increases
103
124
the memory consumption. If you need to set the maximal number of processes,
104
125
the number of processes can be specified with the argument
105
126
``` --num-threads ``` or ``` -n ``` :
@@ -259,13 +280,17 @@ The following table shows the available arguments:
259
280
| -- save-after | Save the intermediate results after a specific number of runs. | 100 |
260
281
261
282
283
+ ---
284
+
262
285
## Notes
263
286
264
287
- If you only have segmentation results and want to evaluate the * SEG* metric,
265
288
it is important that you pass the * --seg* flag to the evaluation command.
266
289
Otherwise, the code could flag your input as invalid, because the
267
290
* res_track.txt* file is missing or inconsistent.
268
291
292
+ ---
293
+
269
294
## Contributing
270
295
271
296
Contributions are welcome! For bug reports or requests please
@@ -280,14 +305,16 @@ pip install pre-commit, pylint
280
305
pre-commit run --all-files
281
306
```
282
307
308
+ ---
309
+
283
310
## Acknowledgement and Citations
284
311
285
312
The code was developed by Timo Kaiser on behalf of the [ Institute of Information
286
313
Processing] ( https://www.tnt.uni-hannover.de/ ) at the Leibniz University Hanover in conjunction with the
287
314
organizers of the [ Cell-Tracking-Challenge] ( https://celltrackingchallenge.net/ ) .
288
315
289
316
290
- If you use this code or the [ CHOTA] ( ) metric in your research, please cite the following paper:
317
+ If you use this code or the CHOTA metric in your research, please cite the following paper:
291
318
``` bibtex
292
319
@inproceedings{kaiser2024chota,
293
320
author = {Kaiser, Timo and Ulman, Vladimír and Rosenhahn, Bodo},
@@ -363,6 +390,8 @@ If you use the MOTA metric in your research, please cite the following paper:
363
390
}
364
391
```
365
392
393
+ ---
394
+
366
395
## License
367
396
368
397
This project is licensed under the BSD 2-Clause License - see the
0 commit comments