Case Study

Computer Vision

Applications in finance, health, and consumer goods

Computer Vision technology infiltrates our daily lives and plays a vital role in many of the most avant-garde B2B and B2C applications. At Qubika’ innovations lab, we’ve successfully integrated Computer Vision technology into finance, health, and consumer goods for prestigious clients such as Google, Nest and Welwaze Medical.

When mapping out a product for potential application of Computer Vision technology, our approach is always user-centric and aimed at an enriched user experience. From custom OCR modules to image processing,
pattern matching and object detection, our Computer Vision integrations have not only made our clients’ products better but oftentimes even possible.

Finance
DTC
UX/UI
iOS
Android

Custom OCR Computer Vision scanning solution integrated into card.io

01

Client

Prepaid2Cash is a mobile iOS & Android solution to convert gift and prepaid cards to easy-to-access cash by sending extracted funds securely to the users’ bank account.

02

Challenge

To ensure the highest level of security when extracting cash out of their gift or prepaid cards, we wanted users to scan their card numbers versus manual input. While there are existing computer vision frameworks for credit card scanning such as card.io, the numbers on gift and prepaid cards are usually not embossed, and thus not detectable.

03

Solution

Our team created a custom Computer Vision solution for Prepaid2Cash, featuring an OCR that is able to recognize non-embossed numbers on prepaid and gift cards. We then embedded our module directly into card.io to leverage the functionalities of this existing framework, enhanced through our custom solution. In addition, we applied 256-bit SSL data encryption and bank-level security to keep user identity and sensitive information safe and secure.

Instead of the laborious and unsafe task of manual input, our custom solution allowed users to quickly and securely scan their card and move to the next steps in a matter of seconds.

Health
DTC
UX/UI
iOS

Pulse estimation through Advanced Computer Vision Analysis

01

Client

In cooperation with next-generation wearable platform Biostrap, PulseCam is a proprietary iOS App that estimates one’s heart rate by pointing the phone’s camera at any face.

02

Challenge

The findings of 2013 Paper “Detecting pulse from head motions in video” (Balakrishnan, Durand, Guttag) spiked our interest in taking Computer Vision pulse detection to iOS with a dynamic camera, which resulted in the development of Pulse Cam.

03

Solution

Our team used advanced computer vision analysis to track changes in skin tone as well as involuntary facial micro-movements caused by the Newtonian reaction of the influx of blood at each heartbeat. We then translated the obtained data into actual heart rate estimates. By using ARKit available on iOS 11, we were then able to display anyone’s pulse in augmented reality above the head, when facing the phone camera.

While Pulse Cam is a fun app to try out with friends and family, the face-to-pulse reading feature could be leveraged for complex products when further developed.

Health
Wearables
Brand
UX/UI
iOS

Detection of breast abnormalities through pattern matching

01

Client

Celbrea offers advanced breast health to women, and consists of a mapping device and an iOS app with advanced dynamic AI algorithms and Blockchain Technology.

02

Challenge

While offering a broad array of health monitoring features such as fertility and birth control which we discussed in detail in our Celbrea Case Study, an essential part of Celbrea is to combat breast cancer by easily detecting any abnormalities in one's breast and feed obtained data into the Celbrea App for analysis and recommendations.

03

Solution

The Celbrea Thermal Mapping Device (TMD) is an FDA 510(k) cleared patch that women can place on their breast to monitor certain abnormalities. Through Computer Vision pattern matching we taught the Celbrea iOS App to recognize the patch and its orientation and detect any thermic color-coded difference between both breasts, indicating potential risk areas. The results can alert one’s physician to the possibility of breast pathology, including occult, thermally active cancer.

While the Celbrea ecosystem approaches women’s health from different angles, Computer Vision technology was essential for the implementation of the breast cancer detection component.