Help the blind see, using tensorflow. Done : Simple integration of tensorflow into an ios app. Use Nuance api to speak to user Design very simple and stupid UI Remove dependence on Nuance. Use Apple or roll my own. Use apple TTS system To Do: Work on creating a better UI. Work on making sure that pipeline is working properly as there are laggs as well as poor accuracy compared to google's code Transfer learn the network on more relevant objects. Currently the model is trained on image net and it is not relevant enough for home usage. Add Face detection and remember feature Add Emotion detection feature Add trafic sign detection feature Reach : Add image to speech conversion feature, so that the blind can read a newspaper.