TESTINT

2 reviews

TestINT (testINT.ai)  What is TestINT? TestINT is a “Data Augmentation and Testing” platform to make systems that utilize “Deep Learning” techniques more reliable. TestINT provides a compact platform for image augmentation & AI testing to help identify AI’s weaknesses. TestINT is specialized for computer vision problems, specifically for street-view and aerial images.  What TestINT does… By applying basic and advanced transformations on your labeled dataset that you use for training and testing, you can replicate your dataset to include situations that are difficult/highly costly to obtain in real life. Thus, you will have a larger, richer and differentiated data set that you can use for testing. Augmentation Features of TestINT: The image augmentation feature of TestINT enables diversifying test data sets through a broad range of image augmentation options in a matter of minutes.  Although new features are continually added, some of the highlights of TestINT are: * TestINT generates new data by applying weather transformations (adding rain, snow, fog),  * Day-to-night image transformation,  * Object placement (car and pedestrian),  * Basic input corruptions (rotation, blurring, contrast, brightness), and  * Adversarial attacks.  TestINT creates challenging conditions synthetically and helps users save time, effort, and money.  How does TestINT support testing? TESTINT’s testing ability enables creating test scenarios by choosing from a variety of model performance and test adequacy metrics. In the end, TestINT prepares insightful test reports for users to identify potential risks of their models. Other Features? Our platform is the first no-code platform that proposes a thorough testing solution for AI-based vision models. TestINT includes 20 augmentation methods for aerial and street-view camera images. TestINT is the only platform with such a broad catalog of image augmentation techniques.  Why should Deep Learning System Developers use TestINT? TestINT lets Deep Learning systems’ developers focus on their core Use Case and model, letting  TestINT provide the dataset to handle different situations. With TestINT, end-products reach the customer better tested, with corner cases handled more comprehensively and more reliably.

Prover

0 reviews

ProVER is a system solution enabling anomaly detection for test engineers monitoring ECUs and their networks for complex vehicles/platforms. Modern vehicles/platforms contain big data and intensive software, and vehicles/platforms under test comprise networks of many ECU (Electronic Control Units) data. ProVER utilizes state-of-the-art data techniques to analyze traces both offline-by importing data from Automotive sectors well-known data logger file formats (or private cloud data) and live data feed from networks and various protocols (i.e. CAN, LIN, FlexRay, Ethernet, UART). Usage and Benefits ProVER, enables software intensive-high data networked automotive/platform developers’ test engineers to analyze and diagnose anomalies more efficiently either at laboratories or at the field by monitoring time-stamped and correlated ECU messages. Reduces Test Engineering efforts and enables easy analysis and detection of anomalies. Enables on-field testing and diagnostic analysis. Enables focusing on anomalies instead of analyzing terabytes’ of data logs. Analysis of anomalies by the correlated data. Monitoring trends of anomalies by graphic user interfaces for prognostic analysis. Features Script Editor for Test Engineers to insert rules Live monitoring of bi-directional flow of ECU data Graphical User Interface to visualize the source of anomalies Live analysis of performance indicators in the form of graphics Visualization for Routing Tables Automatic anomaly detection Integration to existing data recording systems For more information, you can visit our site (www.proven.com.tr), or drop us an email(info@proven.com.tr).