How can I overcome interruption of COD training?

1 次查看（过去 30 天）
Theresa Pflüger
Theresa Pflüger 2020-3-31
编辑： Theresa Pflüger 2020-3-31
Hi Matlab community!
I am currently working with a CascadeObjectDetector to detect objects in images.
The training halts after only 2 stages altough my setting is 5 (desirably I like to have more though). Please find below the screenshot of the warning message.
Consequently, my results and object detections are poor but I guess that is expected from a detector trained based on 2 stages only. How can I improve the training process without interruption from occuring?
I have tried multiple times to adjust the parameters as suggested and added many more negative images as well.
My images contain an object of simple shape on a patterned background, where positive samples contain a single object and negative ones the background only.
The problem persists. I would appreciate any leads to solve this problem!
Thanks in advance!
P.S.: I use MATLAB 2019b on MacOS.

请先登录，再回答此问题。

回答（0 个）

类别

在 Help Center 和 File Exchange 中查找有关 Computer Vision Toolbox 的更多信息

标签

产品


版本

R2019b

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!

Translated by