Why parseGrayscaleImageData and parseRGBImageData of xhrdataset have the same function?

[dsmilkov]

From @cyilu on March 27, 2018 6:41

I think these two function are alike:
`
function parseGrayscaleImageData(
data: Uint8Array|Uint8ClampedArray, result: Uint8Array,
resultOffset: number): void {
let idx = resultOffset;
for (let i = 0; i < data.length; i += 4) {
result[idx++] = data[i];
}
}

function parseRGBImageData(
data: Uint8Array|Uint8ClampedArray, result: Uint8Array,
resultOffset: number): void {
let idx = resultOffset;
for (let i = 0; i < data.length; i += 4) {
result[idx] = data[i];
result[idx + 1] = data[i + 1];
result[idx + 2] = data[i + 2];
idx += 3;
}
}`

Copied from original issue: tensorflow/tfjs-core#902

[dsmilkov]

From @nsthorat on March 27, 2018 14:10

The functions are slightly different, one is parsing a grayscale image (just parsing red channel), and the other is parsing the fully packed RGB byte array. Notice the inner for loop is different.

[dsmilkov]

From @cyilu on March 28, 2018 6:13

@nsthorat But the "data" is 1d array. Therefore, "data" to "result" is 1d array to 1d array. Isn't it？

[dsmilkov]

From @nsthorat on March 28, 2018 15:29

ImageData is always a flat array, you can check out the docs here: https://developer.mozilla.org/en-US/docs/Web/API/ImageData

[dsmilkov]

From @cyilu on March 29, 2018 2:32

@nsthorat Yes, I think these tow function all transfer a flat array to a flat array.

[dsmilkov]

Hi closing this issue since there is an answer posted already. Feel free to reopen if you have additional questions regarding this thread.