System information

• Have I written custom code (as opposed to using a stock example script provided in TensorFlow): No
• OS Platform and Distribution (e.g., Linux Ubuntu 16.04):
  Linux
• Mobile device (e.g. iPhone 8, Pixel 2, Samsung Galaxy) if the issue happens on mobile device:
  NA
• TensorFlow installed from (source or binary):
  Source
• TensorFlow version (use command below):
  1.10
• Python version:
  3.5
• Bazel version (if compiling from source):
  0.15
• GCC/Compiler version (if compiling from source):
  NA
• CUDA/cuDNN version:
  NA
• GPU model and memory:
  NA
• Exact command to reproduce:
  NA

Describe the problem

In the label_image.cc example of TFlite, a resize function present in bitmap_helpers_impl.h is being called in line 171 -
resize<float>(interpreter->typed_tensor<float>(input), in.data(),

In the implementation of the function in bitmap_helpers_impl.h, the following line

  auto output_number_of_pixels =
      wanted_height * wanted_height * wanted_channels;

assumes that for the input image, the height and width must be the same. And this works for the mobilenet example since the input size is 224 x 224. But, this would cause issues for any network where the expected input size does not have the same width and height. It should instead be -

   auto output_number_of_pixels =
      wanted_width * wanted_height * wanted_channels;

Source code / logs

NA