MyFiziq Limited (ASX: MYQ) updated its shareholders with additional performance and data protection enhancements related to MYQ’s technology.
MyFiziq is an advanced technology embedded in the application, utilizing the images from the smartphone to produce a 3D avatar representation with detailed circumference dimensions. MYQ offers B2B services via SDK (Software Development Kit) and white label app to B2B partners across different sectors. The company delivers 97% accurate body measurements with a repeatability of 98%.
The new update from MyFiziq will see it move its patented segmentation model from Amazon’s Web Services platform to Apple’s iOS device platform. The software update allows users to run core functionality directly in their device and addresses several global requirements around transfer and storage of personal data. Currently, the MyFiziq measurement process sends encrypted images to the cloud for processing. The new Image Capture Process SDK will instead use an in-device embedded machine learning model to create non-personally identifiable segmented images for processing. This change allows anonymity to be protected as photos taken never need to leave a user’s device. The new functionality will enable partners to further demonstrate accountability and compliance with data regulations across various regions.
In addition to data privacy benefits, in-device image segmentation also enables other technical and cost-saving advantages. The output of the segmentation process is substantially reduced and compresses the current upload/return time by 95%. This, in turn, allows for multiple attempts to be packaged in single classification object. At approximately 15 seconds per attempt, the end-to-end process is also four times faster than the previous average computes of 60 seconds. Combined, this effectively increases MyFiziq’s throughput capacity 32-fold and results in massive cost reductions for both compute and storage. The new model utilizes Apple’s Metal framework (CoreML) and is part of MyFiziq’s ongoing goal to reduce cloud compute time and cost while maintaining accuracy and repeatability.
The update will allow MyFiziq to process up to 4,800,000 avatar requests an hour without any change to the current Amazon Web Services platform infrastructure MyFiziq has built. An Android version via a TensorFlow model is currently under development and nearing completion.
As per the year-end report on 30 June 2018, MyFiziq incurred an operating loss of $13,919, 468 which includes $12,105, 207 in respect of share-based payments to the suppliers, directors, and employees, under the company’s incentive plans. Even though the company incurred an operating loss, it has a working capital deficit of $460,951 and a net cash outflow from operating activities amounting to $1, 286,211.
Stock performance: MYQ stock has generated a negative return of 9.09 % over the last one month. The shares of MYQ closed the day’s session at A$0.315, up by 5 % or 0.015 points as compared to the last day close of A$0.300. MYQ’s 52-week high and low are marked at A$1.500 and A$0.220 respectively.
This website is a service of Kalkine Media Pty. Ltd. A.C.N. 629 651 672. The website has been prepared for informational purposes only and is not intended to be used as a complete source of information on any particular company. Kalkine Media does not in any way endorse or recommend individuals, products or services that may be discussed on this site. Our publications are NOT a solicitation or recommendation to buy, sell or hold. We are neither licensed nor qualified to provide investment advice.