Super Resolution on Android Phone

• Objective :
  Deploy novel deep learning super-resolution algorithms on smartphones. Initially, the algorithm was trained and tested on a supercomputer. In this work, we tried to deploy the model on a smartphone with a Qualcomm GPU-supported chipset.

• Responsibility:
  1. Deployed a super-resolution on an algorithm on a Qualcomm chipset based android smartphone.
  2. Re-code and re-train an exis ng code to be able to run by the supported libraries.
  3. Use an SNPE library to fully maximize the GPU support unit on a smartphone.

• Demo Code

Github-Repository:

You can visit my repository in https://gitlab.com/aji.setyoko to look the detail of the algorithm. Also, I provide some demo example in case you want to try it.

• Keywords:

#Android Smartphone, #Qualcomm Chipset, #SNPE Library, #Generative Adversarial Network(GAN)