1adrianb / 1adrianb/face-alignment

fail to get 2d landmark in float dtype

Open
#337 3 comments 0 reactions 0 assignees View on GitHub
Dominant language
Python
Stars
7.5k
Forks
1.4k
PR merge metrics
No merged PRs in 30d

Description

I test "fa.get_landmarks_from_image" on some images in 256 * 256 and 512 * 512 resolution. While the result is numpy with dtype float32, but the decimal place of result is always 0, in other words, the dtype of result is int in fact.

Is that a bug? Or how can I solve the problem?

Contributor guide

No contributing guide indexed for this repository

Research direction

Look at the `fa.get_landmarks_from_image` function in the face-alignment library, likely in the main module or API file. Examine how landmarks are processed and returned, focusing on dtype conversion steps. Check if there's a test suite for landmark accuracy to verify the fix. The goal is to ensure returned float32 arrays contain non-zero decimal values.

Written by the indexing model from the issue text.

Assessment

Tech stack
python, pytorch
Domain
ai, computer-vision
Issue type
Bug
Difficulty
3/5
Estimated time
1-2 days
Activity status
Stale
Clarity
Mostly clear
Newbie friendliness
45/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.