MultimodalImageHandler class
Main integration class for handling multimodal image processing in Flutter Gemma to prevent AI image corruption and repeating text pattern issues.
Constructors
Properties
- hashCode → int
-
The hash code for this object.
no setterinherited
- runtimeType → Type
-
A representation of the runtime type of the object.
no setterinherited
Methods
-
noSuchMethod(
Invocation invocation) → dynamic -
Invoked when a nonexistent method or property is accessed.
inherited
-
toString(
) → String -
A string representation of this object.
inherited
Operators
-
operator ==(
Object other) → bool -
The equality operator.
inherited
Static Methods
-
createDiagnosticReport(
{required Uint8List imageBytes, required ModelType modelType, String? processedBase64, String? prompt, String? response}) → DiagnosticReport - Creates a diagnostic report for troubleshooting
-
createMultimodalMessage(
{required String text, required ProcessedImage processedImage, required ModelType modelType, bool isUser = true}) → Message - Creates a properly formatted message for multimodal AI models
-
createTokenizedPrompt(
{required String text, required ProcessedImage processedImage, required ModelType modelType}) → String - Creates a properly tokenized prompt for the AI model
-
extractBase64FromPrompt(
String prompt) → String? - Utility method to safely extract Base64 from various formats
-
processImageForAI(
{required Uint8List imageBytes, required ModelType modelType, String? originalFormat, bool enableValidation = true, bool enableProcessing = true}) → Future< MultimodalImageResult> - Processes and validates an image for use with AI models
-
validateModelResponse(
String response, {required String originalPrompt, required ProcessedImage? processedImage}) → ResponseValidationResult - Validates and handles model responses for corruption patterns