I received a hardware kit at Google Developer Days 2017 India and here is what I made from it.
Let us create a hardware system that classifies different objects. We will interface an NXP i.MX7D with touch screen and camera chip. The objects in front of the camera will be classified and in case of the image being that of a Dog or a Cat, we will further go ahead classifying their breed.
NXP i.MX7D
The i.MX 7Dual delivers high-performance processing for low-power requirements with a high degree of functional integration. The i.MX 7Dual features an advanced implementation of two ARM
Cortex-A7 cores, which operate at speeds of up to 1.2 GHz, as well as the ARM Cortex-M4 core. The Pico variant is pin-compatible with the Intel Edison for sensors and low-speed I/O, but also adds additional expansion possibilities for multimedia and connectivity, giving you cutting edge technology that can easily be expanded and implemented for IoT designs.
Although I have used NXP board, this same project can be carried out using Raspberry Pi too!
Let us have a look at the Project (I made the video at 3:00 AM and my sleep-deprived accent and expressions are something I request you to not judge me by! :-p hahaha)
The results are decent and the model is already available on the internet and doesn't require you to have expertise with TensorFlow.
The only tricky part according to me in this project is interfacing your android mobile phone(iphone not supported, sorry!) with the NXP board and downloading the model on the NXP as this requires a few terminal commands.
Top Comments