2
$\begingroup$

i want to try my hands at a project where i have to detect the brands + the models of sportswears worn by persons participating to a marathon , the project is quite complex , so my first approach would be to use an already pre-trained model that will draw the bounding box of the persons , then use the outputed bonding boxes as input to an other model that will draw me the bounding box for the shirt , shoes , shorts ...ect , and follow this procedure until i get to the brand and then the model of the sportswear. theorically , it's super simple , the problem is that , i'm having a hard time figuring out how to implement it by code , passing the output of one model as an input to an other model , is there any good exemples on the internet can illustrate that approach , by code i mean ?

Any suggestion would be much appreciated , thank you.

$\endgroup$
1
$\begingroup$

I find your question a little hard to understand. If I understand correctly you know how to obtain a bounding box for a person. Using these coordinates, you can select this rectangle using open-cv or matplotlib or any image processing library. If your second model can handle multi-resolution or multiple sizes you can just use this selection as input. If, on the other hand, your second model requires a fixed size you can resize the selection and use padding if necessary.

| improve this answer | |
$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.