Quality Assurance

To sustain a high quality of the Altris AI platform, project goals are defined with the right set of software quality metrics to track them. The quality attributes are defined and continually monitored. This enables the team to ensure final product outputs remain within acceptable, consistent, and high-quality limits along with any introduction of changes, amendments, and new feature implementations.

The following list of quality attributes derives from the needs of the Altris AI platform: 

  1. Functionality: the totality of essential functions of the Altris AI platform is a static attribute and is accumulated in the feature scope. Manual testing of all software items shows that the specified (in requirements documents and business use cases) functionality is met. 

  2. Reliability is being measured in terms of working under different working environments and different conditions for a defined duration. The maximum allowed failure/error rate of the software system (can refer to the entire system or to one or more of its separate functions) for a specified number of input trials under specified input conditions in a specified time interval (assuming that hardware and input are free of errors). A software item can be seen as reliable if this test produces a low error rate (the probability that an error will occur in a specified time interval). 

  3. Usability deals with the staff resources needed to train a new employee and to operate the software system. It also refers to how easily a user can understand the functions of the software and how much effort is required by the users to understand the functions. This can be measured in terms of ease of use. Altris AI platform is user-friendly, easy to learn, and designed in accordance with the UI and UX of the OCT device’s software it’s built into. The applications are easy to use for input preparation (no need to prepare data, it’s already present in the database), operation (one click to initiate the AI analysis), and interpretation of output (clear and simple visualization of pathologies on scans, pathology severities, and report generation function) and are easy for new or infrequent users to learn to use the system. 

  4. Reusability: Altris AI software is highly reusable - different code library classes are generic enough to use easily in different applications. It is designed to be adjusted for each client separately, it also enables future projects to make use of a given module or a group of modules of the currently developed software.

  5. Efficiency of Altris AI software is measured as the ability to fulfill its purpose with the best possible utilization of all necessary resources (time, storage, transmission channels, and peripherals). The upcoming versions are designed (with the user’s consent) to store the depersonalized data of every examination analysis to use the results each time user runs the analysis again. This saves users’ time and the project’s equipment capacity and resources. No hardware resources are spent on the client side. Much effort in the upcoming development is estimated for upgrading of performance and processing speed of AI calculations. 

  6. Portability is measured in terms of technical issues related to porting and behavioral issues related to porting. Considering its architecture, Altris AI software doesn’t need to adapt to hardware environment changes (when any new computer/equipment with OS stated in requirements specification is added to the system), the installation proceeds in a few seconds from one lightweight distributive (executive file). 

  The adaptation takes place when: 

  a. New requirements emerge when the software is used.

  b. The business environment changes.

  c. Repairing errors (changing a system to correct deficiencies in the way it meets its requirements).

  d. the performance or reliability of the system may have to be improved. 

  1. Security and Integrity come together. System integrity and security are sufficient to prevent unauthorized access to system functions, and information loss, and protect the privacy of data entered into the system. 

  2. Learnability. The newly introduced user is able to learn from the design without the tutorial. Nevertheless, the user manual is structured clearly and simply and explains all functionality to the user. 

  3. Flexibility: the software is flexible enough to modify in extremely short terms, though the AI algorithms themselves need reasonable terms to be modified, and these modifications are ongoing. 

  4. Testability: the software is easy to test and find defects, and manual interface and API testing are enough.  

Applying the above quality attributes standards Altris AI determines that the system meets the requirements of quality.