add-grayscale-equalized

[fbtravi]

Added new face detector, the 'face_grayscale_equalized_detector'.

This detector can be used as the 'face_detector', but it is not part of the 'all', this happens so as not to change the use of whom I have the 'all' instantiated.

Issue

[codeclimate]

The PR diff size of 33366 lines exceeds the maximum allowed for the inline comments feature.

[codeclimate]

The PR diff size of 33366 lines exceeds the maximum allowed for the inline comments feature.

[codeclimate]

The PR diff size of 33364 lines exceeds the maximum allowed for the inline comments feature.

[codeclimate]

The PR diff size of 33366 lines exceeds the maximum allowed for the inline comments feature.

[codeclimate]

The PR diff size of 33362 lines exceeds the maximum allowed for the inline comments feature.

[codeclimate]

The PR diff size of 33369 lines exceeds the maximum allowed for the inline comments feature.

[codeclimate]

The PR diff size of 33369 lines exceeds the maximum allowed for the inline comments feature.

[coveralls]

Pull Request Test Coverage Report for Build 3620736147

• 7 of 7 (100.0%) changed or added relevant lines in 2 files are covered.
• No unchanged relevant lines lost coverage.
• Overall coverage increased (+0.4%) to 61.896%

---

Totals
Change from base Build 3433138123:	0.4%
Covered Lines:	277
Relevant Lines:	448

---

💛 - Coveralls

[codeclimate]

The PR diff size of 33430 lines exceeds the maximum allowed for the inline comments feature.

[codeclimate]

The PR diff size of 33428 lines exceeds the maximum allowed for the inline comments feature.

[maiagripp]

LGTM 🎉

[codeclimate]

The PR diff size of 33426 lines exceeds the maximum allowed for the inline comments feature.

[codeclimate]

The PR diff size of 33427 lines exceeds the maximum allowed for the inline comments feature.

[devppjr]

LGTM 🚀

[guilhermef]

@fbtravi, we can add your new detector as part of all, then make it clear on the major release.
Can you address the code climate issues?

[fbtravi]

hi @guilhermef

Yes, I could put it in all, but the doubt would be here.

class CompleteDetector(CascadeLoaderDetector):
    def detect(self, image):
        face_detector = FaceDetector()
        glasses_detector = GlassesDetector()
        profile_detector = ProfileDetector()
        feature_detector = FeatureDetector()

        face_points = face_detector.detect(image) or []
        glasses_points = glasses_detector.detect(image) or []
        profile_points = profile_detector.detect(image) or []
        feature_points = feature_detector.detect(image) or []

        return face_points + glasses_points + profile_points + feature_points

With two face decteres which points do we use? Do we use the points of both?

[guilhermef]

@fbtravi yes, it's not a problem to have overlapping detection unless one of them is giving false positives.

[codeclimate]

The PR diff size of 33470 lines exceeds the maximum allowed for the inline comments feature.

[codeclimate]

The PR diff size of 33477 lines exceeds the maximum allowed for the inline comments feature.

[fbtravi]

Hi @guilhermef, changes made.

[guilhermef]

Awesome work @fbtravi

[RaphaelVRossi]

We should not change year from copyright on old files.

[RaphaelVRossi]

We should not change year from copyright on old files.

[RaphaelVRossi]

We should not change year from copyright on old files.

[heynemann]

Sorry to come late to the party, but I don't think we should merge this. The goal of the entire team of committers should be to reduce thumbor's codebase, not increase it. This detector seems like an awesome addition to the ecosystem, we can even host it in thumbor's org as a separate project, but I'd love if we could trim thumbor's codebase. The more code we have the slower we move. Sorry for taking so long to review this.
That said, I'm ok with merging if you all feel like creating a new project is more trouble than it's worth.

[guilhermef]

@heynemann, given the data presented on #59, shouldn't we make this the default behavior for face detection instead of adding a new detector?
It seems that this would improve the accuracy.

[heynemann]

If that’s the case we can improve our current face detector.

On Sun, 4 Dec 2022 at 06:55 Guilherme Souza ***@***.***> wrote: @heynemann <https://github.com/heynemann>, given the data presented on #59 <#59>, shouldn't we make this the default behavior for face detection instead of adding a new detector? It seems that this would improve the accuracy. — Reply to this email directly, view it on GitHub <#62 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AAAO4JNWLKB6UN6QBRGLKNLWLRS7XANCNFSM6AAAAAASHAFGEI> . You are receiving this because you were mentioned.Message ID: ***@***.***>

-- Bernardo Heynemann

[fbtravi]

The first idea was to create a new detector and sharing almost all the code, not changing the existing one so as not to break any users.
Leaving this unification for a future refactor, but if you think it's better, we can start with this change in the current detector.

If possible, let me know so I can continue with the implementation or not.

@guilhermef @heynemann

[guilhermef]

@fbtravi let's refactor the existing face detection to include your change, then we can release it as a major change.

[codeclimate]

Code Climate has analyzed commit 2e23ddc and detected 0 issues on this pull request.

View more on Code Climate.

[fbtravi]

Would they be able to analyze the changes?

We had a decrease in the lines tested, but this compared to my last commit, the last merge in master is like this

@guilhermef
@heynemann

[RaphaelVRossi]

Great job @fbtravi 🚀