image recognition for mobiles

SHOTPIX is our core technology. It was created to make image recognition a simple network service, a tool to be easily integrated into new generation applications and to support the visual indexing of products and services. The strengths of the SHOTPIX platform can be summarized as follows:

  • Versatility – the service is designed to address the problem of image recognition in a way as much as possible independent from the quality and resolution of images. The technology adopted ensures optimal performance with different light conditions and in the presence of critical geometric transformations like scaling and distortion.
  • Speed – It is a peculiar feature of the SHOTPIX platform, particularly important when aiming at satisfying asynchronous requests coming from mobile devices. Both architectural and computational arrangements ensure fast response of the platform and optimal user experience, at all times.
  • Simplicity – exposing a simple REST API, the platform minimizes the integration time of SHOTPIX in the context of third-party applications (web applications or mobile Apps).


The system features a fully proprietary algorithm based on iconic destructuration of the original image into multiple-class space-variant adaptable features.
This approach makes the image recognition (matching) formally invariant with respect to a large set of offine transformations; recognition can be further enhanced by applying additional statistical and geometric heuristics. In brief, our core technology allows to:

  1.  simplify search and recognition of images having the same iconic content and
  2. operate automatic corrections for rotations, translations, perspective and light variations. The results of the image recognition process are therefore stable within a reasonable range of variations related to the use of “different” and “uncontrolled” (mobile ) acquisition devices.



Like any system, also image recognition systems are prone to errors. These errors are normally described by using two related indexes: false positives (false acceptance rate or FAR) and false negatives (false rejection rate or FRR). FAR represents the percentage of cases in which the system incorrectly identifies an object that is not present in database, confusing it with one of those present, while FRR represents the percentage of cases in which the system does not identify an item which actually exists in the database.
These two values are usually inversely related, and vary depending on the configuration parameters and the recognition threshold adopted. If the system behavior is characterized by a perfect balance (equal percentages) of FAR and FRR, this value is named equal error rate or EER .

In order give a rough idea of the statistical performance of the SHOTPIX platform, internal tests have been designed and operated on medium-sized (>500 images) datasets. These tests give an EER under 12%. Of course, in relation to the number and type of images, the internal thresholds of the SHOTPIX platform can be tuned in order to satisfy specific user needs . For example it is possible to configure the system is such a way to find a matching result with rates close to 100%, but accepting as a counterpart a very high number of errors in the case of elements not present in the database (or vice versa).



The integration with third-party applications is particularly simple because it is based on HTTP requests (WEB SERVICE REST). Client-side integration will therefore be limited to the management of HTTP connections and the coding/decoding of data following the JSON format.

CRUD (create, read, update and delete) management of the image database is guaranteed by a set of additional functionalities , provided through simple APIs. These APIs give the customers a full control on the persistent storage of elements that must be recognized and guarantee smooth integration with the customer’s proprietary platforms