Currently save method works fine for modules that output an RGB image but the current output for modules in bayer raw and YUV do not visually depict the actual result
In the current state of the infinite-ISP Reference Model, there are two noise reduction modules, namely the bilateral filter and 2DNR, which do an excellent job of reducing noise in the image. However, this extra smoothing process tends to remove some fine image details, resulting in a loss of sharpness and overall image quality.
We need to implement a sharpening algorithm into the infinite-ISP reference model to enhance the image quality by retaining important image details while benefiting from the existing modules' noise reduction capabilities.
Our camera ISP pipeline lacks a Contrast Improvement module, resulting in suboptimal image quality, especially in low-light conditions. I propose incorporating a dedicated module to enhance contrast and reveal details in darker regions while preserving overall image quality. This feature will improve the image's contrast and provide a competitive advantage.
An OSD module is essential for various applications, including branding, timestamping, and conveying critical information on video feeds. It will provide users with the flexibility to customize the OSD content and its placement on the screen, enhancing the utility and versatility of our image processing solution.
Currently OSD module is proposed for the purpose of branding. Initial version of OSD is to put 10xEngineers logo on the processed image.