search

weichengkuo / DeepBox

Code release for DeepBox paper in ICCV 2015
Other
128 stars 57 forks source link

Getting results Worse than even Edgebox Proposals #10

Open ghost opened 7 years ago

ghost commented 7 years ago

Hi, I tried your Deepbox implementation with Edgebox proposals and after running it on several images and manually comparing the results of Deepbox to Edgebox, i found that it performed even worse than Edgebox alone. Can you help me where i should improve to get the better result out of it ? What I noticed that scores to those regions are much higher which don't contain objects.

Below you can find image and files containing scores and boxes for Edgebox as well as Deepbox. example.zip

weichengkuo commented 7 years ago

The model was trained on COCO so if you run it on COCO categories it should do better than Edgebox. If you're running it on something else, it'll depend on how similar it is to COCO categories. Hope this helps!

ghost commented 7 years ago

Thanks for the response, I actually want to use it for my own dataset and can you help me with steps need to be taken to train the model for my own dataset and for my application i just need it to be trained on 4 different categories. Any help will be appreciated.