fix: image layout

Signed-off-by: ye jianbai <your_email@abc.example>

Author: ye jianbai

tf-mobilenet-image/README.md

@@ -49,15 +49,15 @@ The `tf` model is located at `./frozen.pb`
 If you want to generate the [raw](birdx224x224x3.rgb) tensor, you can run:
 
 ```shell
-cd rust/image-converter/ && cargo run ../../bird.jpg ../../birdx224x224x3.rgb
+cd rust/image-converter/ && cargo run ../../PurpleGallinule.jpg ../../birdX224X224X3.rgb && cd ../..
 ```
 
 ### Execute
 
 Execute the WASM with the `wasmedge` with Tensorflow Lite supporting:
 
 ```bash
-wasmedge --dir .:. wasmedge-wasinn-example-tf-mobilenet-image.wasm saved_model.pb bird.jpg
+wasmedge --dir .:. wasmedge-wasinn-example-tf-mobilenet-image.wasm saved_model.pb PurpleGallinule.jpg
 ```
 
 You will get the output:

tf-mobilenet-image/birdX224X224X3.rgb

@@ -26,9 +26,9 @@ fn image_to_tensor(path: String, height: u32, width: u32) -> Vec<u8> {
         image::imageops::resize(&img, height, width, ::image::imageops::FilterType::Triangle);
     let mut flat_img: Vec<f32> = Vec::new();
     for rgb in resized.pixels() {
-        flat_img.push(rgb[0] as f32 / 255.);
-        flat_img.push(rgb[1] as f32 / 255.);
-        flat_img.push(rgb[2] as f32 / 255.);
+        flat_img.push(rgb[0] as f32 / 255.0);
+        flat_img.push(rgb[1] as f32 / 255.0);
+        flat_img.push(rgb[2] as f32 / 255.0);
     }
     let bytes_required = flat_img.len() * 4;
     let mut u8_f32_arr: Vec<u8> = vec![0; bytes_required];
@@ -40,7 +40,7 @@ fn image_to_tensor(path: String, height: u32, width: u32) -> Vec<u8> {
             let u8_bytes = u8_f32.to_ne_bytes();
 
             for j in 0..4 {
-                u8_f32_arr[((flat_img.len() / 3 * c + i) * 4) + j] = u8_bytes[j];
+                u8_f32_arr[((i * 3 + c) * 4) + j] = u8_bytes[j];
             }
         }
     }

tf-mobilenet-image/rust/image-converter/src/main.rs

@@ -117,7 +117,7 @@ fn image_to_tensor(path: String, height: u32, width: u32) -> Vec<u8> {
             let u8_bytes = u8_f32.to_ne_bytes();
 
             for j in 0..4 {
-                u8_f32_arr[((flat_img.len() / 3 * c + i) * 4) + j] = u8_bytes[j];
+                u8_f32_arr[((i * 3 + c) * 4) + j] = u8_bytes[j];
             }
         }
     }