Grand Prize Winner

Sergey Konvisarov and his Pencil Music App!

Congratulations Sergey for the creative, innovative and grand-prize winning submission.

See how a high-performance, low power AI solution was brought to life with the on-device mobile machine learning of Snapdragon mobile platforms.

Pencil Music App

The app will generate in real time the band music of different styles from your arbitrary drawing. SNPE accelerated AI will recognize lines, curves, ovals, polylines, their relative displacement, size and other attributes to control the music generator module. It is great fun for kids and adults!

Winners

Devang Mohan

Real time style transfer

Description

An app that can apply style transfer filters on real time camera feed so that the user has the artistic freedom to choose a frame that goes with the style s/he feels during photo composition, instead of the technical restriction of picking a filter that goes with a photo s/he has already taken.

Ognjen Todic

On-device speech recognition

Description

An SDK that provides on-device speech recognition functionality for Snapdragon powered devices.

Bipul Das

On-device AI to deliver mental healthiness

Description

The app will use smartphone front camera to capture facial image streams and apply on-device AI to monitor the underlying emotional state while users use social networking apps or any similar media app. AI algorithm will then help users gain insights of emotional health and provide recommendations.

Hakan Özkelemci

Mobile app for interactive musical instrument training

Description

Mobile app will guide users with step by step interactive tutorials for teaching them to play an instrument(e.g. guitar).App will listen to the users playing the instrument and will give instant feedback on their accuracy as they play along with the interactive tutorial via deep learning techniques.

Ashish Mokalkar

DeepAgro (Smart Agriculture using drones)

Description

DeepAgro captures real-time images of wide area farmlands using drones and allow farmers to easily identify visual crop diseases, predict the ideal height of the crop, recommending possible causes of the disease, suggesting adequate fertilizer usage for higher crop yield.

Eric Kotonya

Artificial Intelligence Rhyming Response Application AIRRAPP

Description

AIRRAPP is a poetic chat mobile app. It lets anyone chat with an AI friend, in a personalized, entertaining way. When you write a line, the AI engine replies with a rhyming sentence, in your language and context.

Slava Vasiliev

Smart Libra

Description

[excuse my English] Traditional weight scales will be improved with 3D camera and ML processing. - Buyer puts any food on the scales surface and system identifies the kind of food - cost and nutritional value for every piece will be estimated. - Info will be stored and sent for further processing.

Navin B

Skin cancer detector

Description

Detecting skin cancer at the earliest is very crucial for effective diagnosis. With machine learning and AI, skin cancer can be detected at early stage. This will help people in rural areas where getting medical examination done is difficult.

Beven Sandengu

Coin-It

Description

Coin-It is a convenient tape measure for your phone. The Augmented Reality app uses coins (as markers) to make precise distance/area measurements of objects in real-time. The app will use Snapdragon NPE to accelerate coin classification, position tracking and automatic calibration.

Gabriel Grand

Real-Time Epileptic Seizure Prediction

Description

Accurate prediction of epileptic seizures requires computationally-intensive time series analysis. By leveraging the Snapdragon NPE, an RNN running on a patent’s mobile device could monitor biometrics through wearable devices to forecast impending seizures minutes or even hours before they occur.

Md. Khairul Alam

Plant Disease Identification

Description

Plant leaves can give a good clue about the disease (e.g. an off-color can be an indication of nutrient). I will develop a mobile app which gives information on plant diseases processing the image of plant leaves taken by the phone camera. It will help farmers to take perfect care of their plant.

Henry Ruiz Guzman

Digital phenotyping tool for crop monitoring

Description

This idea consist in implementing a mobile smart app (free available),that allow to scientists, researchers and farmers, extract physiological information from the image plants data,collected during the crop monitoring process. For economically important food crop (rice, beans and cassava to start).

Francois Lemarchand

Aesthetic Filtering for Photos

Description

This mobile app automatically evaluates the photographs you take based on aesthetic criteria using a deep-neural network. Later, the user can browse through the most beautiful pictures or eventually review all the pictures previously taken in order to tune future aesthetic judgments to their tastes.

Phong Nguyen

Remote marker based tracking for accurate drone landing

Description

The current drone delivery system are based on GPS to perform delivery mission. However,here are problems if the drones operate in urban areas with no GPS signal so it is necessary to build an autonomous drone with accurate landing using integrated processor in the real-time manner.

Francis Walugembe

Baby cry sound detection and analysis algorithm

Description

We are currently developing an artificial intelligence algorithm and software product which has potential for Baby cry sound detection and analysis. Our product will be in position to detect baby cry and at the same time decode its meaning for parents or caregivers alleviate the distress

Sunny aditya

Emma - Dress Helper

Description

Emma, will help people in figuring out clothes which looks good on them.

Pallab Sarkar

NeuralSense: NN based assistant for deaf-mute

Description

NeuralSense will be an assistive application which will empower deaf and mute person to communicate efficiently among themselves and with normal people. This application will also consider the limitation that their is no standardized sign language hence impeding the communication for disabled.

Bassel Ebeed

Image description & Question answering for blind users

Description

An App that can describe an image of a scene from the mobile camera or from the internet(twitter, Facebook ...etc) and can answer speech-based questions related to that scene. this App will be extremely helpful for visually impaired people especially in outdoor scenes and when Browsing the internet.

Deane Landreth

Drive Safe for Travelers

Description

Every year a number of tourists who come to New Zealand die in traffic accidents because they inadvertently drive on the wrong side of the road. Traffic here drives on the left. This application will monitor the road ahead and warn if it detects a vehicle is on the wrong side of the road.

Resources

Get Inspired. To help you gather all the insight, information and direction you need, we have provided links to the following resources:

Qualcomm Developer Network is a community of coders, makers and thought-leaders. We’re working together to share ideas and inspiration with YOU and the next generation of high-performance experiences and designs.

Learn more

TensorFlow is an an open-source machine learning platform for everyone. It's a fast, flexible and production-ready framework that scales from research to production.

Learn more

Caffe is a deep learning framework made with expression, speed, and modularity in mind. Take a look.

Learn more

Caffe2 provides an easy way for you to experiment with deep learning. You can bring your creations to scale using the power of GPUs in the cloud or to the masses on mobile with Caffe2’s cross-platform libraries.

Learn more

About the AI Developer Contest

To enter the AI Developer Contest, describe in English and in 600 characters or less, what YOU would create with on-device mobile machine learning and artificial intelligence on a Snapdragon mobile platform. An optional photo may be included.

You have until October 13, 2017 to submit your entry. Throughout the contest all eligible entries will be posted on the webpage (developforai.com/submissions) for others to see. Entries will be judged based on creativity 25%, best potential use of the Snapdragon processing power 25%, and real-world application 50%.

Twenty (20) winners will be announced on October 20, 2017 and each will receive the prize of a Snapdragon 835 Mobile Hardware Development Kit. These 20 winners are eligible to compete for the grand prize trip for two to Mobile World Congress in Barcelona, by submitting an APK of their demo along with a supporting video bringing their idea to life, by no later than December 8, 2017.

Contest rules