feat: some improvements

Author: NorbertKlockiewicz

apps/computer-vision/app/object_detection/index.tsx

@@ -50,49 +50,6 @@ export default function ObjectDetectionScreen() {
     }
   };
 
-  const runForwardPixels = async () => {
-    try {
-      console.log('Testing with hardcoded pixel data...');
-
-      // Create a simple 320x320 test image (all zeros - black image)
-      // In a real scenario, you would load actual image pixel data here
-      const width = 320;
-      const height = 320;
-      const channels = 3; // RGB
-
-      // Create a black image (you can replace this with actual pixel data)
-      const rgbData = new Uint8Array(width * height * channels);
-
-      // Optionally, add some test pattern (e.g., white square in center)
-      for (let y = 100; y < 220; y++) {
-        for (let x = 100; x < 220; x++) {
-          const idx = (y * width + x) * 3;
-          rgbData[idx + 0] = 255; // R
-          rgbData[idx + 1] = 255; // G
-          rgbData[idx + 2] = 255; // B
-        }
-      }
-
-      const pixelData: PixelData = {
-        dataPtr: rgbData,
-        sizes: [height, width, channels],
-        scalarType: ScalarType.BYTE,
-      };
-
-      console.log('Running forward with hardcoded pixel data...', {
-        sizes: pixelData.sizes,
-        dataSize: pixelData.dataPtr.byteLength,
-      });
-
-      // Run inference using unified forward() API
-      const output = await ssdLite.forward(pixelData, 0.3);
-      console.log('Pixel data result:', output.length, 'detections');
-      setResults(output);
-    } catch (e) {
-      console.error('Error in runForwardPixels:', e);
-    }
-  };
-
   if (!rfDetr.isReady) {
     return (
       <Spinner

packages/react-native-executorch/common/rnexecutorch/data_processing/ImageProcessing.cpp

@@ -225,7 +225,7 @@ readImageToTensor(const std::string &path,
   if (tensorDims.size() < 2) {
     char errorMessage[100];
     std::snprintf(errorMessage, sizeof(errorMessage),
-                  "Unexpected tensor size, expected at least 2 dimentions "
+                  "Unexpected tensor size, expected at least 2 dimensions "
                   "but got: %zu.",
                   tensorDims.size());
     throw RnExecutorchError(RnExecutorchErrorCode::UnexpectedNumInputs,

packages/react-native-executorch/common/rnexecutorch/utils/FrameExtractor.cpp

@@ -88,8 +88,7 @@ cv::Mat extractFromAHardwareBuffer(void *hardwareBuffer) {
                             errorMessage);
   }
 
-  // Note: We don't unlock here - Vision Camera manages the lifecycle
-
+  AHardwareBuffer_unlock(buffer, nullptr);
   return mat;
 #else
   throw RnExecutorchError(RnExecutorchErrorCode::PlatformNotSupported,
@@ -111,4 +110,4 @@ cv::Mat extractFromNativeBuffer(uint64_t bufferPtr) {
 #endif
 }
 
-} // namespace rnexecutorch::utils
+} // namespace rnexecutorch::utils

packages/react-native-executorch/common/rnexecutorch/utils/FrameExtractor.h

@@ -22,4 +22,4 @@ namespace rnexecutorch::utils {
  */
 cv::Mat extractFromNativeBuffer(uint64_t bufferPtr);
 
-} // namespace rnexecutorch::utils
+} // namespace rnexecutorch::utils

packages/react-native-executorch/common/rnexecutorch/utils/FrameProcessor.cpp

@@ -25,4 +25,4 @@ cv::Mat extractFrame(jsi::Runtime &runtime, const jsi::Object &frameData) {
 
   return extractFromNativeBuffer(bufferPtr);
 }
-} // namespace rnexecutorch::utils
+} // namespace rnexecutorch::utils

packages/react-native-executorch/common/rnexecutorch/utils/FrameProcessor.h

@@ -24,4 +24,4 @@ using namespace facebook;
  */
 cv::Mat extractFrame(jsi::Runtime &runtime, const jsi::Object &frameData);
 
-} // namespace rnexecutorch::utils
+} // namespace rnexecutorch::utils

packages/react-native-executorch/src/controllers/BaseOCRController.ts

@@ -2,24 +2,11 @@ import { Logger } from '../common/Logger';
 import { symbols } from '../constants/ocr/symbols';
 import { RnExecutorchErrorCode } from '../errors/ErrorCodes';
 import { RnExecutorchError, parseUnknownError } from '../errors/errorUtils';
-import { Frame, PixelData, ResourceSource, ScalarType } from '../types/common';
+import { isPixelData } from '../modules/computer_vision/VisionModule';
+import { Frame, PixelData, ResourceSource } from '../types/common';
 import { OCRLanguage, OCRDetection } from '../types/ocr';
 import { ResourceFetcher } from '../utils/ResourceFetcher';
 
-function isPixelData(input: unknown): input is PixelData {
-  return (
-    typeof input === 'object' &&
-    input !== null &&
-    'dataPtr' in input &&
-    input.dataPtr instanceof Uint8Array &&
-    'sizes' in input &&
-    Array.isArray(input.sizes) &&
-    input.sizes.length === 3 &&
-    'scalarType' in input &&
-    input.scalarType === ScalarType.BYTE
-  );
-}
-
 export abstract class BaseOCRController {
   protected nativeModule: any;
   public isReady: boolean = false;

packages/react-native-executorch/src/hooks/computer_vision/useImageSegmentation.ts

@@ -1,3 +1,4 @@
+import { PixelData } from '../..';
 import {
   SemanticSegmentationModule,
   SegmentationLabels,
@@ -49,7 +50,7 @@ export const useSemanticSegmentation = <
   });
 
   const forward = <K extends keyof SegmentationLabels<ModelNameOf<C>>>(
-    imageSource: string,
+    imageSource: string | PixelData,
     classesOfInterest: K[] = [],
     resizeToInput: boolean = true
   ) =>

packages/react-native-executorch/src/hooks/computer_vision/useSemanticSegmentation.ts

@@ -55,6 +55,8 @@ export function useModuleFactory<
 
     return () => {
       currentInstance?.delete();
+      setInstance(null);
+      setIsReady(false);
     };
 
     // eslint-disable-next-line react-hooks/exhaustive-deps