maddo/AIFotoOnlus
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages 1 Wiki Activity Actions

Compare commits

base: maddo:master
Branches Tags
maddo:master
maddo:0.1.3
maddo:0.1.2
maddo:0.1.1
maddo:v0.1.0
maddo:0.1.0
maddo:v0.1.0.1
maddo:v0.1.0.0
..
compare: maddo:0.1.2
Branches Tags
maddo:master
maddo:0.1.3
maddo:0.1.2
maddo:0.1.1
maddo:v0.1.0
maddo:0.1.0
maddo:v0.1.0.1
maddo:v0.1.0.0

No commits in common. "master" and "0.1.2" have entirely different histories.
master ... 0.1.2

3 changed files with 30 additions and 107 deletions
Download patch file Download diff file Expand all files Collapse all files

6
src/AIFotoONLUS.Core/AIFotoONLUS.Core.xml
View file

@ -139,12 +139,6 @@
must match the class ordering used by the trained recognition network. must match the class ordering used by the trained recognition network.
</summary> </summary>
</member> </member>
<member name="P:AIFotoONLUS.Core.ModelConfiguration.UseGpu">
<summary>
When enabled, request OpenCV DNN CUDA backend/target for inference.
The installed OpenCV runtime must have CUDA support or model loading/forwarding may fail.
</summary>
</member>
<member name="P:AIFotoONLUS.Core.ModelConfiguration.EnableCropSaving"> <member name="P:AIFotoONLUS.Core.ModelConfiguration.EnableCropSaving">
<summary> <summary>
When enabled, recognition crops will be saved to disk under When enabled, recognition crops will be saved to disk under

6
src/AIFotoONLUS.Core/ModelConfiguration.cs
View file

@ -55,12 +55,6 @@ namespace AIFotoONLUS.Core
/// </summary> /// </summary>
public string[] NumberClasses { get; set; } = new[] { "0", "1", "2", "3", "4", "5", "6", "7", "8", "9" }; public string[] NumberClasses { get; set; } = new[] { "0", "1", "2", "3", "4", "5", "6", "7", "8", "9" };
/// <summary>
/// When enabled, request OpenCV DNN CUDA backend/target for inference.
/// The installed OpenCV runtime must have CUDA support or model loading/forwarding may fail.
/// </summary>
public bool UseGpu { get; set; } = false;
/// <summary> /// <summary>
/// When enabled, recognition crops will be saved to disk under /// When enabled, recognition crops will be saved to disk under
/// "logs/crops" for diagnostic inspection. Disabled by default. /// "logs/crops" for diagnostic inspection. Disabled by default.

125
src/AIFotoONLUS.Core/NumberRecognitionEngine.cs
View file

@ -95,8 +95,10 @@ namespace AIFotoONLUS.Core
_detectionNet = CvDnn.ReadNetFromDarknet(_cfg.DetectionCfg, _cfg.DetectionWeights); _detectionNet = CvDnn.ReadNetFromDarknet(_cfg.DetectionCfg, _cfg.DetectionWeights);
_recognitionNet = CvDnn.ReadNetFromDarknet(_cfg.RecognitionCfg, _cfg.RecognitionWeights); _recognitionNet = CvDnn.ReadNetFromDarknet(_cfg.RecognitionCfg, _cfg.RecognitionWeights);
ConfigureNetRuntime(_detectionNet, _cfg.UseGpu); _detectionNet.SetPreferableBackend(Backend.OPENCV);
ConfigureNetRuntime(_recognitionNet, _cfg.UseGpu); _detectionNet.SetPreferableTarget(Target.CPU);
_recognitionNet.SetPreferableBackend(Backend.OPENCV);
_recognitionNet.SetPreferableTarget(Target.CPU);
// Let OpenCV use multiple threads internally (use number of logical processors) // Let OpenCV use multiple threads internally (use number of logical processors)
try try
{ {
@ -106,11 +108,6 @@ namespace AIFotoONLUS.Core
{ {
// Ignore if not supported by OpenCvSharp build // Ignore if not supported by OpenCvSharp build
} }
if (_cfg.UseGpu)
{
ValidateGpuRuntime();
}
} }
public void Dispose() public void Dispose()
@ -122,38 +119,6 @@ namespace AIFotoONLUS.Core
GC.SuppressFinalize(this); GC.SuppressFinalize(this);
} }
public static bool TryValidateGpuRuntime(ModelConfiguration cfg, ILogger? logger, out string? failureMessage)
{
if (cfg is null) throw new ArgumentNullException(nameof(cfg));
var probeConfiguration = new ModelConfiguration
{
DetectionCfg = cfg.DetectionCfg,
DetectionWeights = cfg.DetectionWeights,
RecognitionCfg = cfg.RecognitionCfg,
RecognitionWeights = cfg.RecognitionWeights,
ConfidenceThreshold = cfg.ConfidenceThreshold,
NmsThreshold = cfg.NmsThreshold,
DetectionInputSize = cfg.DetectionInputSize,
RecognitionInputSize = cfg.RecognitionInputSize,
NumberClasses = cfg.NumberClasses,
EnableCropSaving = cfg.EnableCropSaving,
UseGpu = true
};
try
{
using var engine = new NumberRecognitionEngine(probeConfiguration, logger);
failureMessage = null;
return true;
}
catch (Exception ex)
{
failureMessage = ex.GetBaseException().Message;
return false;
}
}
private static string SanitizeFileName(string name) private static string SanitizeFileName(string name)
{ {
foreach (var c in Path.GetInvalidFileNameChars()) name = name.Replace(c, '_'); foreach (var c in Path.GetInvalidFileNameChars()) name = name.Replace(c, '_');
@ -162,39 +127,6 @@ namespace AIFotoONLUS.Core
private string[] GetOutputLayerNames(Net net) => net.GetUnconnectedOutLayersNames(); private string[] GetOutputLayerNames(Net net) => net.GetUnconnectedOutLayersNames();
private static void ConfigureNetRuntime(Net net, bool useGpu)
{
if (useGpu)
{
net.SetPreferableBackend(Backend.CUDA);
net.SetPreferableTarget(Target.CUDA);
return;
}
net.SetPreferableBackend(Backend.OPENCV);
net.SetPreferableTarget(Target.CPU);
}
private void ValidateGpuRuntime()
{
try
{
using var detectionProbe = new Mat(_cfg.DetectionInputSize.Height, _cfg.DetectionInputSize.Width, MatType.CV_8UC3, Scalar.All(0));
_ = DetectTextRegions(_detectionNet, detectionProbe).Take(1).ToArray();
using var recognitionProbe = new Mat(_cfg.RecognitionInputSize.Height, _cfg.RecognitionInputSize.Width, MatType.CV_8UC3, Scalar.All(0));
using var blob = CvDnn.BlobFromImage(recognitionProbe, 0.00392, _cfg.RecognitionInputSize, new Scalar(0, 0, 0), true, false);
_recognitionNet.SetInput(blob);
using var output = _recognitionNet.Forward();
}
catch (Exception ex)
{
throw new InvalidOperationException(
"OpenCV DNN CUDA runtime validation failed. Disable number AI GPU mode or use an OpenCV runtime built with CUDA DNN support.",
ex);
}
}
/// <summary> /// <summary>
/// Detect text regions in the supplied image using the detection network. /// Detect text regions in the supplied image using the detection network.
/// </summary> /// </summary>
@ -220,7 +152,7 @@ namespace AIFotoONLUS.Core
var outNames = GetOutputLayerNames(detectionNet); var outNames = GetOutputLayerNames(detectionNet);
var outsList = new List<Mat>(); var outsList = new List<Mat>();
detectionNet.Forward(outsList, outNames); detectionNet.Forward(outsList, outNames);
Mat[] outs = outsList.ToArray(); Mat[] outs = outsList.ToArray();
if (outs.Length == 0) if (outs.Length == 0)
{ {
@ -230,15 +162,15 @@ namespace AIFotoONLUS.Core
var fallback = new List<Mat>(); var fallback = new List<Mat>();
for (int on = 0; on < outNames.Length; on++) for (int on = 0; on < outNames.Length; on++)
{ {
try try
{ {
var single = detectionNet.Forward(outNames[on]); var single = detectionNet.Forward(outNames[on]);
fallback.Add(single); fallback.Add(single);
} }
catch (Exception ex) catch (Exception ex)
{ {
_logger?.LogError(ex, "Fallback Forward failed for {name}", outNames[on]); _logger?.LogError(ex, "Fallback Forward failed for {name}", outNames[on]);
} }
} }
if (fallback.Count > 0) if (fallback.Count > 0)
{ {
@ -289,21 +221,21 @@ namespace AIFotoONLUS.Core
} }
if (maxScore > _cfg.ConfidenceThreshold) if (maxScore > _cfg.ConfidenceThreshold)
{ {
int x = (int)Math.Max(0, Math.Round(cx - w / 2)); int x = (int)Math.Max(0, Math.Round(cx - w / 2));
int y = (int)Math.Max(0, Math.Round(cy - h / 2)); int y = (int)Math.Max(0, Math.Round(cy - h / 2));
var rect = new Rect(x, y, (int)Math.Round(w), (int)Math.Round(h)); var rect = new Rect(x, y, (int)Math.Round(w), (int)Math.Round(h));
boxes.Add(rect); boxes.Add(rect);
confidences.Add(maxScore); confidences.Add(maxScore);
classIds.Add(bestClass); classIds.Add(bestClass);
centerXList.Add(cx); centerXList.Add(cx);
} }
} }
} }
if (boxes.Count == 0) return Enumerable.Empty<DetectedRegion>(); if (boxes.Count == 0) return Enumerable.Empty<DetectedRegion>();
CvDnn.NMSBoxes(boxes, confidences, (float)_cfg.ConfidenceThreshold, (float)_cfg.NmsThreshold, out int[] indices); CvDnn.NMSBoxes(boxes, confidences, (float)_cfg.ConfidenceThreshold, (float)_cfg.NmsThreshold, out int[] indices);
@ -554,8 +486,10 @@ namespace AIFotoONLUS.Core
{ {
var det = CvDnn.ReadNetFromDarknet(_cfg.DetectionCfg, _cfg.DetectionWeights); var det = CvDnn.ReadNetFromDarknet(_cfg.DetectionCfg, _cfg.DetectionWeights);
var rec = CvDnn.ReadNetFromDarknet(_cfg.RecognitionCfg, _cfg.RecognitionWeights); var rec = CvDnn.ReadNetFromDarknet(_cfg.RecognitionCfg, _cfg.RecognitionWeights);
ConfigureNetRuntime(det, _cfg.UseGpu); det.SetPreferableBackend(Backend.OPENCV);
ConfigureNetRuntime(rec, _cfg.UseGpu); det.SetPreferableTarget(Target.CPU);
rec.SetPreferableBackend(Backend.OPENCV);
rec.SetPreferableTarget(Target.CPU);
netsBag.Add((det, rec)); netsBag.Add((det, rec));
return (det, rec); return (det, rec);
}); });
@ -591,7 +525,8 @@ namespace AIFotoONLUS.Core
try try
{ {
using var tempRec = CvDnn.ReadNetFromDarknet(_cfg.RecognitionCfg, _cfg.RecognitionWeights); using var tempRec = CvDnn.ReadNetFromDarknet(_cfg.RecognitionCfg, _cfg.RecognitionWeights);
ConfigureNetRuntime(tempRec, _cfg.UseGpu); tempRec.SetPreferableBackend(Backend.OPENCV);
tempRec.SetPreferableTarget(Target.CPU);
var alt = RecognizeDigits(crop, tempRec, ctx); var alt = RecognizeDigits(crop, tempRec, ctx);
if (!string.IsNullOrEmpty(alt)) txt = alt; if (!string.IsNullOrEmpty(alt)) txt = alt;
} }