🐛 Bug Report: focal point is wrong for GIFs whose first frame does not start at (0, 0)
Evaluación
Este issue todavía no se ha evaluado.
Descripción
👟 Reproduction steps
A GIF has a logical screen size, and each frame has its own position inside it. When the first frame covers only part of the screen (for example an 800×450 frame at x=800 on a 1600×450 screen), /analyze returns gravity measured against the frame instead of the whole image.
imageutil.Decode checks the logical screen size with image.DecodeConfig. It then returns the frame that image/gif decodes, and that frame keeps the frame's own bounds ((800,0)-(1600,450) here). Everything after that (face detection, saliency, FocalNet crops) normalizes against the frame's width and height. The result is then read as a point on the full image.
- Build a GIF with a 1600×450 logical screen and one frame,
internal/testimages/testdata/generated-face-left.jpg(800×450), placed at left=800, top=0. Any encoder that writes a frame offset works, for example Go'sgif.EncodeAllwithConfig{Width: 1600, Height: 450}and a frame whose bounds areimage.Rect(800, 0, 1600, 450). curl -s -H 'Content-Type: image/gif' --data-binary @face-left-offset.gif http://127.0.0.1:8080/analyze
👍 Expected behavior
The face is at x ≈ 0.26 within the 800 px frame. On the 1600 px image that is x ≈ (800 + 0.26 × 800) / 1600 ≈ 0.63.
👎 Actual Behavior
{"gravity":{"x":0.26323946937918663,"y":0.23157380289501617},"confidence":0.9248611985862869,"source":"face"}
x = 0.26 is in the empty left half of the image. FocalNet crop pixel coordinates have the same problem, because they are computed from img.Bounds().Dx()/Dy() of the frame.
🎲 Appwrite version
autogravity main @ 35b155f
💻 Operating system
macOS (arm64)
🧱 Your Environment
Built from source with Go 1.26.1 and ONNX Runtime 1.30.0, using the default U²-Net + YuNet backend with the checked-in INT8 models.
👀 Have you spent some time to check if this issue has been raised before?
- I checked and didn't find similar issue
🏢 Have you read the Code of Conduct?
- I have read the Code of Conduct
I have a fix with a regression test ready. Could this be assigned to me so I can open a PR?
- Lenguaje dominante
- Go
- Estrellas
- 4
- Forks
- 5
- Merge medio
- 8 h 54 min
- PR fusionados (30 d)
- 3
Preparar el entorno
- Incluye un Dockerfile o un archivo de Docker Compose
- Sin plantilla de pull request
- Leer la guía de contribución
Primeros pasos
- Lee el issue completo y luego la guía de contribución del proyecto.
- Comenta en el issue que vas a ocuparte — evita que dos personas hagan lo mismo.
- Haz un fork del repositorio y trabaja en una rama.
- Abre un pull request que haga referencia al número del issue.
Más de appwrite/autogravity
-
Dependency DashboardAbierto
Dificultad 4/5 3-5 días Aptitud para principiantes 8/100
appwrite/autogravity#24 ·
Todos los issues de appwrite/autogravity
Issues similares
-
Dificultad 2/5 1-3 horas Aptitud para principiantes 76/100
OneBusAway/maglev#2145 ·
Los mantenedores suelen responder en 2 días
-
priority/medium status/ready type/chore
Dificultad 2/5 1-3 horas Aptitud para principiantes 84/100
discobox-ai/discobox#1357 ·
Los mantenedores suelen responder en 1 día
-
TestPseudoLocaleRemote fails now and then: it types the code before the laptop asks for itPosiblemente ocupada @fzn0x la tomó hoy. Abiertobug priority: low
Dificultad 2/5 1-3 horas Aptitud para principiantes 88/100
Los mantenedores suelen responder en 1 día
-
enhancement ui/ux
Dificultad 2/5 1-3 horas Aptitud para principiantes 68/100
jbigler/railswitch#243 ·
Los mantenedores suelen responder en 1 día
-
go-vulnerability security-scan
Dificultad 1/5 Menos de una hora Aptitud para principiantes 84/100
teslashibe/agent-go#35 ·