Research and Engineering

productHunt

github

Yashesh Bharti  • Peerlist

<h3>PulseProof offers a revolutionary approach to verifying human presence by leveraging unique physiological signals. Utilizing the heartbeat as a form of proof, PulseProof ensures that only real humans can access services, effectively combating bots and deepfakes. The system prioritizes user privacy by processing all sensitive data, including raw heartbeat signals, directly on the user's device. This on-device processing means that no raw health data is ever shared with servers, ensuring a high level of privacy and security.</h3><ul><li>Core Functionality: PulseProof transforms a live heartbeat signal, captured via an iPhone camera or Apple Watch, into a verifiable cryptographic proof. This proof is then used to confirm the presence of a live human.</li><li>Privacy-Centric Design: The product is built with privacy at its core. Raw heartbeat data is processed locally on the device, and only a signed verification token is sent to the application. This ensures 0 bytes of health data are shared.</li><li>Trust Tiers: PulseProof offers multiple verification tiers to suit different needs:<ul><li>Tier 1 (Camera): A lightweight check using the iPhone camera, ideal for bot resistance in scenarios like account creation or waitlists.</li><li>Tier 2 (Watch + Camera): Combines two live signals for stronger confidence, suitable for sensitive logins or payments.</li><li>Tier 3 (Sovereign): A full private proof flow with zero-knowledge verification, where the system verifies the cryptographic proof without accessing biometric data.</li></ul></li><li>Developer Integration: PulseProof provides a drop-in SDK for iOS, Android, and Web applications, allowing for quick integration in as little as 5 minutes. Developers receive a signed token upon successful verification, which can then be validated on their backend.</li><li>Security Model: The system separates liveness verification from biometric collection, focusing on local signal processing, short-lived proof tokens, and auditable verification results.</li><li>Use Cases: PulseProof is ideal for moments where human presence is critical, including bot-resistant logins, payment confirmations, account recovery, marketplace trust, AI platform integrity, and high-risk private presence checks.</li></ul>

PulseProof of Humanity

COVID-19 Litmus Test which uses State of the Art Machine learning model to determine if the Chest X-Ray that you upload is COVID-19 positive or not within seconds. The model has 99.6% accuracy. Datasets sourced from COVID Chest XRAY dataset for COVID-19 infected lungs and Kaggle Pneumonia XRAY Dataset for healthy lungs. This development could prove vital for countries where the population is extremely high, and if we consider a large portion of the population being affected even if we perform 100,000 testes per day it would take years to test all of them and practicality of producing 100,000 test kits is also diminishingly small. With time could achieve 100% accuracy Programmed with Keras, Tensorflow, OpenCV  and Python

COVID-19 Detection using Deep Learning

There are softwares which translate from any language to any language, however, the promise of the futuristic sign language decoders have always been far away from being a day to day reality. Designing a Neural Network based Sign language to speech translator which purports to translate sign language in real-time to text or speech as the user gestures. For people in the Deaf community, this tool could come in handy as it will not demand any extra hardware and use a camera (Like the ones present on the phone) to detect the pose and gestures. Using Tensorflow, Openpose, OpenCV, and Python