TDA4VM: void PostprocessImageObjDetect::operator()(void frameData, VecDlTensorPtr &results)

Shine

Part Number: TDA4VM

Hello Champs,

How does the void *PostprocessImageObjDetect::operator()(void *frameData, VecDlTensorPtr &results) function in post_process_image_object_detect.cpp get the output data of the object detection module? Does it get the data from the results variable? How to analyze it?

Thanks
Regards,
Shine

over 2 years ago

0 Takuma Fujiwara over 2 years ago

TI__Guru 50403 points

Hi Shine,

I would recommend analyzing the function from the bottom up. Essentially, we are trying to get the following values from the results variable to overlay boxes on top of frameData:

numEntries - to determine how many objects are detected
score - to determine if we want to overlay box or ignore the detected object due to low confidence
label - to determine what class we are detecting
box - to determine top left and bottom right coordinates to know where to draw box

Regards,

Takuma

0 Shine over 2 years ago in reply to Takuma Fujiwara

TI__Genius 13655 points

Hi Takuma,

Thanks for your great support.

Customer wanted to add the post process result. It need to get the data from the output layer of the detection model. How to get it from the results?

Is the following code correct?

auto *f_tensor = reinterpret_cast<float *>(results[0]->data);

Is f_tensor a pointer which points to the output layer data? If yes, what is the data type and dimension?

Thanks
Regards,
Shine

0 Takuma Fujiwara over 2 years ago in reply to Shine

TI__Guru 50403 points

Hi Shine,

The logic that we have in our code to get the output is within the getVal function: https://git.ti.com/cgit/edgeai/edge_ai_apps/tree/apps_cpp/common/src/post_process_image_object_detect.cpp#n166

I would recommend using similar, if not the same function to extract the output layer.

Regards,

Takuma

Processors

Processors forum

TDA4VM: void PostprocessImageObjDetect::operator()(void frameData, VecDlTensorPtr &results)

Processors

Processors forum

TDA4VM: void *PostprocessImageObjDetect::operator()(void *frameData, VecDlTensorPtr &results)

TDA4VM: void PostprocessImageObjDetect::operator()(void frameData, VecDlTensorPtr &results)