Real-time image processing on passthrough imagery?

Yes, you're correct that the Vision framework in iOS primarily focuses on high-level tasks like object recognition and image analysis. However, for classic image processing operations like spatial filtering, histogram equalization, and image warping, you might want to consider using the Core Image framework.

Core Image is a powerful framework in iOS that allows you to apply various image processing filters and operations to images and videos in real-time. It provides a wide range of built-in filters and tools that you can use for tasks like color correction, blurring, sharpening, and more.

You can use Core Image to create custom filter chains to perform operations like spatial filtering, histogram equalization, and image warping. The framework is designed to take advantage of hardware acceleration, making it suitable for real-time processing tasks.

Here's a general outline of how you might approach this using Core Image:

1. Creating a CIContext: You would create a CIContext object that specifies the rendering options, such as whether to use CPU or GPU acceleration.

2. Loading the Image: You can load the incoming camera frames as CIImage objects. These images can then be processed using the Core Image filters.

3. Applying Filters: You can use various Core Image filters to perform operations like spatial filtering, histogram equalization, and image warping. For spatial filtering, you can use filters like CIGaussianBlur, CISharpenLuminance, etc. For histogram equalization, you might need to create a custom filter chain that involves modifying the pixel values.

4. Displaying the Processed Image: After applying the desired filters and operations, you can render the processed CIImage to the screen using the CIContext.

While Core Image is a versatile framework for image processing, keep in mind that the specific implementation details will depend on your application's requirements and the desired effects you're aiming to achieve. It's also important to consider performance implications when working with real-time processing, as some filters might be computationally intensive.

To sum up, Core Image could be a suitable option for implementing classic image processing operations in real-time from incoming camera imagery before displaying them on OLED screens.