Hi, First of all thank you so much for this tutorial. I followed all your steps but I am getting framerate of only 10 FPS. Can you please specify any reasons if possible. I am using RPI4B with 4GB ram.

*. Code : MobileNetV1.cpp
....
cout << "tensors size: " << interpreter->tensors_size() << "\n";
cout << "nodes size: " << interpreter->nodes_size() << "\n";
cout << "inputs: " << interpreter->inputs().size() << "\n";
cout << "input(0) name: " << interpreter->GetInputName(0) << "\n";
cout << "outputs: " << interpreter->outputs().size() << "\n";
....

*. The output is something like this.
use : detect.tflite
tensors size: 184
nodes size: 64
inputs: 1
input(0) name: normalized_input_image_tensor
outputs: 4
Fps : 28.5714

*. How I can change tensors_size, nodes size, outputs?
use : mobilenet_v1_1.0_224_quant.tflite
tensors size: 90
nodes size: 31
inputs: 1
input(0) name: input
outputs: 1
Fps : 41.4583

use : mobilenet_v2_1.0_224_quant.tflite
tensors size: 174
nodes size: 65
inputs: 1
input(0) name: input
outputs: 1
Fps : 50

*. I want to increase my FPS in Mobilenet V1, V2

Hi

I am in desperate need of a fast Hand-Detection model.
Although I wasn't able to reach the 24 fps (I am stuck at around 10 with overclocking, 7 without - loading the model-file in python), this is still the fastest detection I got to run on my Pi 4.

But its for the "wrong" objects - I need only the detection of hands.

While comparing your model with a pretrained PyTorch implementation of the ssd mobile net v3 I noticed that even the official models are too slow. Even on an 2023 MacBook I reached only 5 fps.

Could you point me in the somewhat right direction for how to customize the model for my own use case?

I was thinking of using the EgoHands dataset

Ff my request is considered inappropriate here, feel free to delete it without comment.

Best regards,
Jan

Get this error using Ubuntu 20 or RaspiOS 64

/usr/bin/ld: /root/tensorflow/tensorflow/lite/tools/make/gen/linux_aarch64/lib/libtensorflow-lite.a(interpreter_builder.o)||undefined reference to symbol 'dlsym@@GLIBC_2.17'|

How is the model file trained？And how to train my own dataset？
thanks

Hi,

We have a application to protect my friends hands . running on normal PC 25fps

what would be the best model and AI platform to detect hand like below to prevent accidents in RPi ? should be fast as possible as

Camera fixed up and see directly
thx

i want to ignore some objects like refrigator. How can i edit COCO_labels.txt file? I try to change label name with none, null, unlabeled, but when i run code, just refrigator name changed with none, null or unlabeled

I'm running this example, but it has a problem. When capturing camera V2, the input frames are CROPPED automatically into 640x480. This is really terrible, I want the full angle of the camrera, and rescale to 300x300 before put into the SSDv1 model.
I read the videoio.hpp but had no idea.
Please, help me to figure out directions to resolve it.
Thank you so much.