demo

[mahilaMoghadami]

hello
thank you for implementation sharing
I want to test model on video or image. did you share your model? if yes, which file is for testing and inference?
thank you.

[Cli98]

Hi @mahilaMoghadami

Thank you for your interest! We host models (MCNN + object detection)via provided pretrain links, which have been provided via google drives. Please follow the step in my description to inference your image.

Thank you and let me know if you have any further questions.

[mahilaMoghadami]

Hi @mahilaMoghadami

Thank you for your interest! We host models (MCNN + object detection)via provided pretrain links, which have been provided via google drives. Please follow the step in my description to inference your image.

Thank you and let me know if you have any further questions.

thank you @Cli98
how I can use this architecture as end-to-end network?
I mean that: I want to get output of MCNN (for generating density maps) then crop regions and then fed this patchs to detector. but I don't know how?

imagine I want to reimplement it and train on visdrone dataset again.
appreciate if help.
thank you

[Cli98]

Hi @mahilaMoghadami
Thank you for your interest! We host models (MCNN + object detection)via provided pretrain links, which have been provided via google drives. Please follow the step in my description to inference your image.
Thank you and let me know if you have any further questions.

thank you @Cli98 how I can use this architecture as end-to-end network? I mean that: I want to get output of MCNN (for generating density maps) then crop regions and then fed this patchs to detector. but I don't know how?

imagine I want to reimplement it and train on visdrone dataset again. appreciate if help. thank you

Hi @mahilaMoghadami

Following the steps here to make it end-to-end, in case you wanna to reimplement it.

1. Run MCNN to get density crops. You can find it at here: https://github.com/CommissarMa/MCNN-pytorch, and pretrain weights here here
2. Run code in image-cropping folder to generate image crops.
3. for the crops+original image, use a state-of-the-art detector to detect , then run code in "fusion detection" to get final detection result.

In step #2, you can run this command to generate density crops:
python density_slide_window_official.py . HEIGHT_WIDTH THRESHOLD --output_folder Output_FolderName --mode val
Please replace all constant (in upper letter) with yours.

Let me know if you need any more help.