Create threshold.ts

[BadMachine]

Image processing operation threshold with few methods.

To see the logs from the Cloud Build CI, please join either our discussion or announcement mailing list.

---

This change is 

[google-cla]

Thanks for your pull request. It looks like this may be your first contribution to a Google open source project (if not, look below for help). Before we can look at your pull request, you'll need to sign a Contributor License Agreement (CLA).

📝 Please visit https://cla.developers.google.com/ to sign.

Once you've signed (or fixed any issues), please reply here with @googlebot I signed it! and we'll verify it.

---

What to do if you already signed the CLA

Individual signers

• It's possible we don't have your GitHub username or you're using a different email address on your commit. Check your existing CLA data and verify that your email is set on your git commits.

Corporate signers

• Your company has a Point of Contact who decides which employees are authorized to participate. Ask your POC to be added to the group of authorized contributors. If you don't know who your Point of Contact is, direct the Google project maintainer to go/cla#troubleshoot (Public version).
• The email used to register you as an authorized contributor must be the email used for the Git commit. Check your existing CLA data and verify that your email is set on your git commits.
• The email used to register you as an authorized contributor must also be attached to your GitHub account.

ℹ️ Googlers: Go here for more info.

[BadMachine]

@googlebot I signed it!

[lina128]

Thank you for the contribution, @BadMachine! I left some comments, see below.

Reviewable status: 0 of 1 approvals obtained (waiting on @BadMachine)

---

tfjs-core/src/ops/image/threshold.ts, line 1 at r4 (raw file):

import { Tensor3D } from '../../tensor';

Add license.

/**
 * @license
 * Copyright 2021 Google LLC. All Rights Reserved.
 * Licensed under the Apache License, Version 2.0 (the "License");
 * you may not use this file except in compliance with the License.
 * You may obtain a copy of the License at
 *
 * https://www.apache.org/licenses/LICENSE-2.0
 *
 * Unless required by applicable law or agreed to in writing, software
 * distributed under the License is distributed on an "AS IS" BASIS,
 * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 * See the License for the specific language governing permissions and
 * limitations under the License.
 * =============================================================================
 */

---

tfjs-core/src/ops/image/threshold.ts, line 9 at r4 (raw file):

/**
 * Performs threshold algorithms on Tensors

Sugg: Performs image thresholding, which creates a binary image from a grayscale image.

---

tfjs-core/src/ops/image/threshold.ts, line 10 at r4 (raw file):

/**
 * Performs threshold algorithms on Tensors
 * @param image a 2d image tensor of shape `[x , y, 3]`.

Sugg: 3d tensor of shape [batch, imageHeight, imageWidth, depth]. The image color range should be [0, 255].

---

tfjs-core/src/ops/image/threshold.ts, line 12 at r4 (raw file):

 * @param image a 2d image tensor of shape `[x , y, 3]`.
 * @param method Optional string from `'binary' | 'otsu'`,
 *     defaults to binary, which specifies type of the threshold operation

Sugg: which specifies the method for thresholding.

---

tfjs-core/src/ops/image/threshold.ts, line 13 at r4 (raw file):

 * @param method Optional string from `'binary' | 'otsu'`,
 *     defaults to binary, which specifies type of the threshold operation
 * @param coeff Optional number which defines thresh coefficient from 0 to 1.

Is there a default?

---

tfjs-core/src/ops/image/threshold.ts, line 14 at r4 (raw file):

 *     defaults to binary, which specifies type of the threshold operation
 * @param coeff Optional number which defines thresh coefficient from 0 to 1.
 * @param inverted Optional boolian which

boolean, which specifies if color should be inverted. Defaults to false.

---

tfjs-core/src/ops/image/threshold.ts, line 56 at r4 (raw file):

}

function threshold_(

Move the doccomment next to this method.

---

tfjs-core/src/ops/image/threshold.ts, line 59 at r4 (raw file):

  image: Tensor3D | TensorLike,
  method: 'binary' | 'otsu' = 'binary',
  coeff?: number,

What's the default?

---

tfjs-core/src/ops/image/threshold.ts, line 60 at r4 (raw file):

  method: 'binary' | 'otsu' = 'binary',
  coeff?: number,
  inverted?: boolean

What's the default?

---

tfjs-core/src/ops/image/threshold.ts, line 63 at r4 (raw file):

): Tensor3D {
    
  const $image = convertToTensor(image, 'image', 'threshold');

Do you also need to check dtype?

---

tfjs-core/src/ops/image/threshold.ts, line 65 at r4 (raw file):

  const $image = convertToTensor(image, 'image', 'threshold');
  const threshold = coeff * 255;
  const arrayedImage = Array.from($image.dataSync());

This is a blocking call. Why not use tfjs ops? See suggestion on how to use it below.

---

tfjs-core/src/ops/image/threshold.ts, line 74 at r4 (raw file):

  util.assert(
    method === 'otsu' || method === 'binary',
    () => `method must be binary or otsu, but was ${method}`);

Capital M.

---

tfjs-core/src/ops/image/threshold.ts, line 76 at r4 (raw file):

    () => `method must be binary or otsu, but was ${method}`);

  if (method === 'binary') {

const condition = inverted ? less($image, threshold) : greater($image, threshold);
const result = where(condition, 255, 0);
return result;

---

tfjs-core/src/ops/image/threshold.ts, line 86 at r4 (raw file):

  } else if (method === 'otsu') {
    const histogram = Array(256).fill(0);

Maybe you can use tfjs.bincount for getting the histogram?

---

tfjs-core/src/ops/image/threshold.ts, line 95 at r4 (raw file):

    const threshold = otsu_alg(histogram, arrayedImage.length);

    arrayedImage.forEach((item, i) => {

What about inverted?

---

tfjs-core/src/ops/image/threshold.ts, line 96 at r4 (raw file):

    arrayedImage.forEach((item, i) => {
      arrayedImage[i] = (item < threshold) ? 0 : 255;

Can you use tfjs ops same as the basic case.

[lina128]

Reviewable status: 0 of 1 approvals obtained (waiting on @BadMachine)

---

tfjs-core/src/ops/ops.ts, line 303 at r7 (raw file):

// Second level exports.
export {image, linalg, losses, spectral, fused, signal, sparse};

Why is sparse removed?

---

tfjs-core/src/ops/image/threshold.ts, line 96 at r7 (raw file):

        grayscale = add(add($r, $g), $b);
    } else {
        grayscale = image;

What is this condition for? When color channel is 4, what happens? If the method doesn't support the case, it's better to throw an error early on.

---

tfjs-core/src/ops/image/threshold_test.ts, line 23 at r7 (raw file):

describeWithFlags('threshold', ALL_ENVS, () => {
	it('default method binary, no arguments passed but input image', async () => {

Why is there '>>' sign?

[lina128]

Hi @BadMachine , once you addressed the three comments, I think the PR is good to go, great work!

Reviewable status: 0 of 1 approvals obtained (waiting on @BadMachine)

[BadMachine]

@lina128 Thanks for your attention!

Why is sparse removed?

Have no idea, didn't do anything except importing threshold op.

Why is there '>>' sign?

Excuse me, what did you mean by '>>' sign?

[lina128]

Reviewable status: 0 of 1 approvals obtained (waiting on @BadMachine)

---

tfjs-core/src/ops/ops.ts, line 300 at r8 (raw file):

import {sparseReshape} from './sparse/sparse_reshape';
const sparse = {sparseReshape};

This two lines are still removed. Please add back.

[lina128]

This is what I see in Reviewable UI, see the '<<' signs, please reformat:

[BadMachine]

@lina128
Gotcha!

[lina128]

Reviewable status: 0 of 1 approvals obtained (waiting on @BadMachine)

---

tfjs-core/src/ops/image/threshold_test.ts, line 23 at r7 (raw file):

Previously, lina128 (Na Li) wrote…

Why is there '>>' sign?

I think you use 2 tabs vs. 2 spaces.

[BadMachine]

@lina128

What is this condition for? When color channel is 4, what happens? If the method doesn't support the case, it's better to throw an error early on.

Channels check has been added.

    util.assert(
        $image.shape[2] === 3 || $image.shape[2]=== 1,
        () => 'Error in threshold: ' +
            'image color channel must be equal to 3 or 1' +
            `but got ${$image.shape[2]}.`);

[BadMachine]

@lina128

Also added threshold_test.ts
Extended /ops/ops.ts
Imported './ops/image/threshold_test' to tests.ts

But for some reason building failed with ERROR: build step 10 "node:10" failed: step exited with non-zero status: 1

[lina128]

Reviewable status: 0 of 1 approvals obtained (waiting on @BadMachine)

---

tfjs-core/src/ops/image/threshold.ts, line 130 at r10 (raw file):

    for (let index = 0; index < histogram.size-1; index++) {

        classFirst = histogram.slice(0, index + 1);

Remove chained op.

---

tfjs-core/src/ops/image/threshold.ts, line 132 at r10 (raw file):

        classFirst = histogram.slice(0, index + 1);

        classSecond = histogram.slice(index + 1);

Remove chained op.

[BadMachine]

@lina128

Remove chained op.

Done.

[lina128]

Thank you for your contribution, @BadMachine! LGTM!

Reviewable status: complete! 1 of 1 approvals obtained (waiting on @BadMachine)

[BadMachine]

@lina128

Yippee! )