Chef.ai is a web app that was created to simplify the 'What should I eat?' process. It combines a myriad of technologies, including the Intel Distribution of OpenVINO, to make figuring out what to cook a fun, simple, process.
This project builds a predictive model to forecast an onsetting epileptic seizure before 5 minutes. The two main models used for prediction are xgboost (gradient-based decision tree) and convolutional neural networks. This work achieves a ROC-AUC of 0.79.
Voicenet is a comprehensive library for performing speech/voice-based functions. It is capable of doing:
Speech to text (STT)
Gender detection based on the voice.
Pronunciation posterior score
Articulation-rate
Speech rate
Filler words
Age detection from voice.
Emotion detection from voice.
This application can predict different stages of Diabetic Retinopathy with a decent accuracy. The reason behind this research work is to create a cheap and reliable DR detection solution which can be easily accessed irrespective of proper internet connectivity.
The proposed deep network connection is used over state-of-the-art MobileNet-V2 architecture and manifests two cases, which lead from 33.5% reduction in flops (one connection) up to 43.6% reduction in flops (two connections) with minimal impact on accuracy.
A tennis ball retriever using #Intel 's D415 RealSense camera and #OpenVINO for object detection. Great project to use computer vision, object detection, and robotics!
The hotdog app of Intelligent Robotics.
Learning to Rank is the problem involved with ranking a sequence of documents based on their relevance to a given query. In this work, we propose DeepQRank, a deep q-learning approach to this problem.
We set up google football environment for testing A3C RL algorithm for the same. For the implementation of RL algorithms, we have used ChainerRL library given it contains an optimized version of A3C. We have used one file namely a3c.py which contains the code for training the agent.
This project is based on the online Dataset - 'Mnist Dataset' consisting of 10,000 images of Alphabets/Letters on which Image classification has been applied to recognize the letters.
It is always important to understand how the visitor feels in a retail store. But, no visitor is in any mood to give written feedback or even to press buttons. Hence, we came up with this innovative approach where they just have to show emotions! We record the feedback!
This project uses vision to read data matrix from real-time. Data Matrix is a two-dimensional barcode consisting of black and white "cells" or modules arranged in either a square or rectangular pattern, also known as a matrix.
This CNN Model is used to detect if an input image of a character is rotated or not. If it is rotated, it returns the rotation angle so that the rotation can be inverted and it can be classified using the standard CNN like ResNet, Inception, etc.
In this project I have made a deep learning generative model which uses a segmented land cover map and generates a imagery map out of it. It utilizes the latest SPADE normalization layer in the generative model to not which keeps the context of input land cover map.