Infinite-ISP
Infinite-ISP copied to clipboard
A camera ISP (image signal processor) pipeline that contains modules with simple to complex algorithms implemented at the application level.
2DNR should be placed before sharpen module, which is empty currently, so that final image get both better signa- noise-ratio and sharpness.
So currently the the pipeline can run on only a single image. Enable this pipeline to run on multiple images using a single run/execution
The BNR algorithm just like DPC also takes a lot of time for its execution. When this module is turned on the complete pipeline takes a lot of time to...
The deal pixel correction or defected pixel correction takes a lot of time to execute. This is mainly because it uses loops that slows down the execution. Implementation this algorithm...
Till now all parameters are not being checked for validity in most of the modules. A check needs to be added before giving them to the algorithm.
For now the image is converted to 8 bits after CFA and further in the pipeline 8bit range is used.
The repo supports the conversion of YUV to 4:4:4 and 4:2:2 and saves them on a .yuv file based on the selection made in the config file. A reader function...
When I sync the repo it says **8 commits ahead of xx-isp/infinite-isp:main**. This is weird as both the repo when compared have zero changes except for their commit hashes.
Grey world white balancing has already been implemented in [white_balance.py](https://github.com/xx-isp/infinite-isp/blob/main/modules/white_balance.py). This is at Bayer domain. A similar algorithm for white world white balancing is required.