Real-time estimation of Heart Rate from facial images under different lighting conditions using Smartphone Camera
Smart phone application that measures heartrate using on-board cameras in real-time and non-invasively offers several advantages and have multiple use cases. Detecting heart rate using only camera has an added advantage because in such case the users do not require any skills or compliance, but has to just capture image/video. During exercise, it is often desired that the user can measure their heart rate without any extra accessory. Wearing a wearable during exercise can be uncomfortable to certain users. Catering to the phenomenon of people using camera as the most popular smartphone feature we can address this problem.
This work will predominantly focus on estimating Heart Rate in real-time from facial images captured using smartphone camera. With every Heart Beat blood flows to our face. Whenever our heart circulates blood, the amount of light reflected from the face varies. To the human eye, this variation is invisible. But using image processing techniques we can detect these changes. The underlying principle is to analyse the spatio-temporal variations in the time-series facial images and amplify these variations. The time series of color values at pixels in region of interest is used to estimate the heart rate. For this work data needs to be collected at various lighting conditions: i) Daytime Outdoor (Cloudy and Sunny) ii) Daytime Indoor iii) Night-time Indoor. Data needs to be acquired for resting and exercising condition. This work broadly requires work in the domain of video processing, signal processing and basic concepts of machine learning.
Goal: Reduce carbon footprint of the travel
Proposal: Build a smartphone app that helps you to car/bike pool, compare and compete with your friends on the carbon footprint of the travel.
Air pollution in India is a serious issue with major contributors being vehicle emissions and traffic congestions. Individuals and communities can play a more effective role in reducing the air pollution if they are aware of their contributions to the same. Carbon footprint is one way of quantifying the direct and indirect emissions.
We propose smartphone application (app) that can automatically track the carbon footprint of the users by analysing the details of the commute, such as mode of travel, distance, and duration.
Basic guidelines of the app:
1. Community centred – the app lets you invite family and friends to join a community that strives to reduce the carbon footprint
2. Gamification – Virtual points/badges awarded to the members who top the eco-friendliness of commute, or make significant improvements to their carbon footprint
3. Simple to use – the app may request the user to setup their regular mode of travel (private vehicle or public transport) during the onboarding time. The app infers other details of the commute automatically through smartphone sensors (GPS etc.) and related framework, for e.g., Android’s DetectedActivity class.
4. Recommendations – the app encourages pooling by dividing the carbon footprint among the pool participants. In addition, it recommends the public transport for individuals by providing the timing of public transport.
Considering privacy concerns event based sensors such as DVS (dynamic vision sensor) are preferable in smart home than normal RGB cameras. In addition to that DVS is power and data efficient compared to RGB cameras. But DVS captures events corresponds to object under motions which does not contain any texture information like RGB cameras. The problem is recognition of daily essential activities of humans like walking, bathing, eating, cooking .etc using DVS video clips and provide analytics to corresponding services. Domains Involved: Image processing, Computer vision, Machine learning
Problem Statement Type
19th December , 2018
19th December , 2018
Your Application is not approved. Please contact to Admin .
Problem statements from Organizations/ MinistriesStudents Innovation
Select Registration Type
Problem Statement Creater College Faculty SPOC Students Team Leader