JieHe96 / iOS_Tensorflow_ObjectDetection_Example

An iOS application of Tensorflow Object Detection with different models: SSD with Mobilenet, SSD with InceptionV2, Faster-RCNN-resnet101

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

three images for tf.session->run

powderluv opened this issue · comments

Do you know why we send in three images as the input ?

out_tensors->push_back(resized_tensor);
out_tensors->push_back(image_tensors_org);
out_tensors->push_back(image_tensors_org4);

Seems like we send tensors of shape:
image_tensors[0]: Tensor<type: float shape: [1,224,224,3]
image_tensors[1]: Tensor<type: uint8 shape: [1,900,1352,3]
image_tensors[2]: Tensor<type: uint8 shape: [1,900,1352,4]

though according to summarize_graph
Found 1 possible inputs: (name=image_tensor, type=uint8(4), shape=[1,?,?,3])