Choosing a face recognition technology solution that is right for your project can be a daunting task. These are some areas that differentiate piXserve from other products in the market.
| piXserve | Competitors | |
|---|---|---|
| Level of technology | +✔ | ~✔ |
| Level of Productization | Complete Product with UI | API only |
| Deployment Options | Cloud and On-Premises | Cloud only |
| Cost | Lower Cost | Higher cost |
| Data protection | Fully Controlled by Customer | Not transparent/Not known |
| Technical Support | Telephone/On-line/Maintenance Contract | On-line |
The piXserve face detection and recognition capabilities are state-of-the-art. piXserve distinguishes itself in difficult situations, when faces are presented in a range of lighting situations, face rotation, face size, image quality, etc. piXserve succeeds where other systems fail. piXserve is designed to operate at scale. piXserve works with images and videos in over 100 formats, as well as live video sources. Customers come to appreciate piXserve after trying competing internet-based offerings.
Most competing face recognition products are offered as a web service, accessible via an API. To make use of the API, the customer needs to apply skilled engineering resources who can build and debug a pipeline and a product that in turn its users can access and work with. This can be expensive and time consuming. In contrast, piXserve is a complete prepackaged solution with an extensive set of features and workflows that help increase user productivity. With a few clicks of the mouse and an intuitive UI, users can create databases, search databases, automatically add face names to databases, and perform many other operations common in many workflows and use case scenarios. No developer resources required, and your staff can be productive immediately. Face recognition is only one of the features available in piXserve. Visual search, object detection and autotagging, text recognition, and other features are standard and can help you refine your face searches to help you get to the particular "needle in the haystack" that if important to you. This level of ease of use and integration into a complete environment is unparalleled in the industry.
piXserve can be deployed on-premises or on the Cloud. The application is fully self contained, so it can run on a server at your office/data-center, or as a virtual machine on the Cloud (Private or Public). In either case, the application is fully controlled by the customer at all times. piXserve does not need to access anything externally in order to operate. piXserve scales "horizontally" so that as workloads increase they can be accomodated by adding physical or virtual servers, automatically communicating with each other. As a result, mixed modes and federated modes are also supported (a combination of physical and/or virtual servers connected via TCP/IP VPN tunnels/encryption). In many situations, customers work with data that resides and can only be accessed in private networks. The ability to deploy piXserve on premises is a key feature for these customers.
Competing face recognition products offered on the web come with a "per-image" pricing structure. This price only covers the cost of generating face metadata. Other costs apply if the customer wishes to create references databases and other ancillary components. In the case of video processing, costs escalate very quickly. In contrast, piXserve is sold on the cloud by the hour, regardless of the number of images or videos that are put through the machine during the time period. The savings from using piXserve become significant very rapidly because an instance of piXserve can process a large quantity of images/videos in a short period of time (depending on the number of cores/GPUs selected by the customer for running the instance). At the low end, when the throughput volume requirement is low, the economics are not important because the level of expenditures is low. As the volume increases to half a million, a million, or more, the piXserve advantage becomes large and the work can be done at a fraction of the cost of alternatives. When deployed on premises, piXserve is available both on a perpetual licensee basis and on a yearly subscription basis. The cost is per server, 2 CPUs/server, unlimited number of cores, unlimited number of clients, no limits on the amount of images/videos that can be processed (other than as permitted by the physical characteristics of the underlying server). For customers that work with continuing flows of images/videos, piXserve is a very cost effective solution.
Face recognition services on the web require that the customer upload their images/video files to be processed on servers owned and operated by the service provider. There is no transparency to the process. The customer does not know what happened to the file after it was uploaded: where was the image sent on the vendor's internal network, where was it processed, what happened to the image after processing, was the image deleted fully, was it cached, was it used for unspecified vendor internal processes, what happened to the data that was returned to the user from the API call, was it deleted fully, was it cached, was it used for other purposes, etc. Many customers work with sensitive data and are required to keep full control of every aspect of data handling. They need to be sure that the data has not been altered or repurposed, that it has been properly erased, and that unauthorized eyes have not access to it whether in transit or in the vendor shop. In contrast, piXserve places the customer in total and absolute control. piXserve runs as a fully self contained application on a physical or virtual server. No external connections or processing is required. Only the customer is privy to the data that piXserve is asked to process, as well as any output, log, database, or anything associated with the work done, The customer decides and controls everything about the process and its inputs and outputs.
Most web-based face recognition services provide self-service, on-line technical support. Primarily the support is geared towards developers and usage of the APIs. piXserve is supported by real people that one can actually speak to and correspond with. What is covered is not only developer support, but also system administration support and user support. When working with applications of face recognition technologies that are expected to operate at-scale, a little human hand-holding can save significant time and reduce operating costs.