dotnet/machinelearning · error · InvalidOperationException

Object is disposed.

Error message

Object is disposed.

What it means

MLImage wraps a disposable SKBitmap; ThrowInvalidOperationExceptionIfDisposed is called at the start of most members (PixelFormat, GetBGRPixels, Pixels, Tag, Width, Height, Save, etc.). Once Dispose ran, _image is null and any member access throws 'Object is disposed.'

Solutions

1. Do not call members after Dispose; keep the image alive for its entire usage scope
2. Re-load the image (MLImage.CreateFromFile) if it was disposed prematurely
3. Wrap access in try-catch for InvalidOperationException and treat as a lifecycle bug
4. Check ObjectDisposed-like state via try/catch since MLImage exposes no IsDisposed property

Example fix

// before
using var image = MLImage.CreateFromFile(path);
DoWork(image);
// disposed here, then:
var w = image.Width; // throws
// after
MLImage image = MLImage.CreateFromFile(path);
var w = image.Width;
DoWork(image);
image.Dispose();

// no public IsDisposed; track ownership yourself
bool usable = image != null && !_ownsAndDisposedSet.Contains(image);

static bool IsUsable(MLImage img) => img is not null; // track Dispose calls in your own wrapper

try { var w = image.Width; } catch (InvalidOperationException ex) when (ex.Message == "Object is disposed.") { image = MLImage.CreateFromFile(path); }

When it happens

Trigger: Accessing any property or method on an MLImage after calling Dispose() — including Width, Height, Pixels, Tag, PixelFormat, GetBGRPixels, or Save.

Common situations: Using images inside ML.NET pipelines after an owning using-block disposed them; double-dispose followed by access; caching MLImage instances across requests while disposal happens concurrently.

Understand the failure class

Background: "Invalid state transition" errors: "status must be X, actually Y", "already rejected/charging/uninstalled", "cannot ... while running" — what they mean when a library rejects your call — this error's family across 31 libraries.

Related errors

• Couldn't resize the image
• Invalid imagePixelData buffer size.
• Invalid resize mode value.
• Path has invalid image file extension.
• Pixel data is unavailable.

AI-assisted analysis of dotnet/machinelearning@7b76e69cf9 (2026-09-11). Data as JSON: /api/errors/f98dae707a6ca43c. Report an issue: GitHub.

Appendix: source

Thrown at src/Microsoft.ML.ImageAnalytics/MLImage.cs:452

            { ".jpg", SKEncodedImageFormat.Jpeg },
            { ".jpeg", SKEncodedImageFormat.Jpeg },
            { ".gif", SKEncodedImageFormat.Gif },
            { ".ico", SKEncodedImageFormat.Ico },
            { ".astc", SKEncodedImageFormat.Astc },
            { ".avif", SKEncodedImageFormat.Avif },
            { ".dng", SKEncodedImageFormat.Dng },
            { ".heif", SKEncodedImageFormat.Heif },
            { ".ktx", SKEncodedImageFormat.Ktx },
            { ".pkm", SKEncodedImageFormat.Pkm },
            { ".wbmp", SKEncodedImageFormat.Wbmp },
            { ".webp", SKEncodedImageFormat.Webp }
        };

        private void ThrowInvalidOperationExceptionIfDisposed()
        {
            if (_image is null)
            {
                throw new InvalidOperationException("Object is disposed.");
            }
        }
    }

    /// <summary>
    /// Specifies the format of the color data for each pixel in the image.
    /// </summary>
    public enum MLPixelFormat
    {
        /// <summary>
        /// The pixel format is unknown.
        /// </summary>
        Unknown,

        /// <summary>
        /// Specifies that the format is 32 bits per pixel; 8 bits each are used for the blue, green, red, and alpha components.
        /// The color components are stored in blue, green, red, and alpha order
        /// </summary>

View on GitHub (pinned to 7b76e69cf9)