You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Add multi-points input, foreground/background points input and box input to EfficientSAM model (#291)
* a
* add efficientsam model and basic demo
* update license
* remove example images
* update readme
* update readme
* update demo
* update demo
* update readme
* update SAM and __init__
* update demo and sam
* update label
* add present gif
* update readme
* add efficientSAM gif to readme of opencvzoo
* cv version 4.10.0, remove camera branch
* 1. add multipoints infering(max: 6)
2. add box prompt(drag), add background point(long press)
3. model fix to 1024*1024
4. label padding -1
5. update demo
* replace the model by new model support mutil-points input, update demo
* update readme
* update readme
* change window size to (800*600), pictures be put in can not exceed it
* add int8 model
* update demo
* update README
* check OpenCV version
* update model name in demo
* update model name in demo
* Add a key to exit ('q' and 'Q'); When clicks reach maximum, no box shows; comment useless print, delete useless whitespace
* update demo with some ASCII
Copy file name to clipboardExpand all lines: models/image_segmentation_efficientsam/README.md
+13-5Lines changed: 13 additions & 5 deletions
Original file line number
Diff line number
Diff line change
@@ -3,9 +3,16 @@
3
3
EfficientSAM: Leveraged Masked Image Pretraining for Efficient Segment Anything
4
4
5
5
Notes:
6
-
- The current implementation of the EfficientSAM demo uses the EfficientSAM-Ti model, which is specifically tailored for scenarios requiring higher speed and lightweight.
7
-
- MD5 value of "efficient_sam_vitt.pt" is 7A804DA508F30EFC59EC06711C8DCD62
8
-
- SHA-256 value of "efficient_sam_vitt.pt" is DFF858B19600A46461CBB7DE98F796B23A7A888D9F5E34C0B033F7D6EB9E4E6A
6
+
- The current implementation of the EfficientSAM demo uses the EfficientSAM-Ti model, which is specifically tailored for scenarios requiring higher speed and lightweight.
7
+
- image_segmentation_efficientsam_ti_2024may.onnx(supports only single point infering)
@@ -17,7 +24,7 @@ Run the following command to try the demo:
17
24
python demo.py --input /path/to/image
18
25
```
19
26
20
-
Click only **once** on the object you wish to segment in the displayed image. After the click, the segmentation result will be shown in a new window.
27
+
**Click** to select foreground points, **drag**to use box to select and **long press** to select background points on the object you wish to segment in the displayed image. After clicking the **Enter**, the segmentation result will be shown in a new window. Clicking the **Backspace** to clear all the prompts.
21
28
22
29
## Result
23
30
@@ -41,4 +48,5 @@ All files in this directory are licensed under [Apache 2.0 License](./LICENSE).
0 commit comments