📜 ⬆️ ⬇️

Scientific corporate Smart Engines (or as we went to ICMV 2018)


Every self-respecting IT company appreciates its employees. This is expressed not only in good salaries, flexible schedules and home comfort in the office (which is undoubtedly an important material indicator of employee value), but also in the conduct of various events from the “for the soul” series. That is why the news often flashes notes about how large IT companies rent restaurants / guest houses / yachts for various kinds of corporate events. And, probably, this is all very cool, friendly and fun. Probably, in such moments, the same mythical team building, about which people love to talk abroad, is actively thriving. And nevertheless, we in Smart Engines have our own view on how to spend honestly earned money with the greatest benefit for employees.


Dear reader, if you accidentally stumbled upon this article and looked under the cat, then let me introduce myself. We are a team of Smart Engines, engaged in the recognition of digital images on an industrial scale. Our main focus is the recognition of documents on devices of various formats: from budget mobile phones to large servers (including domestic production).


Having already created the best product for recognizing Russian passport images, we do not stop at what we have achieved and move forward. Therefore, having brought together all our thoughtfulness, perseverance and perseverance, adding a few romantic views, we progressively go forward, moving along with the science in part of the recognition of digital images.


And this approach to business automatically left an imprint on our creative team. We are one of the few successful IT companies where all employees are scientists. We are proud of it.


As a result, the publication of articles in leading scientific journals, participation in world-wide conferences, the defense of candidate and doctoral studies is our work. And this work concerns not only fulltime employees, but also students and postgraduates studying at our department. And as a reward for the good work, Smart Engines fully assumes all the costs of a trip to the conference. Our employees, students and graduate students, in the course of their work, annually travel to Europe and Asia with reports, combining scientific activities with personal recreation (we all understand that at any self-respecting conference there is always an extensive cultural program).


While practicing to speak at conferences, the guys are pumping up their professional level in several areas at once: interpersonal skills, a foreign language, narration logic, the ability to convey a basic message, etc. And such skills, as well as the ability to swim, are difficult to acquire and almost impossible to lose. And of course, all these skills will be useful on the protection of diplomas and candidates, at interviews and, of course, subsequent scientific presentations.


And now a few photos from our last conference - The 11th International Conference on Machine Vision (ICMV 2018) . From Smart Engines, 15 people participated in it, among them about half were our undergraduates.



Section “Session A-3: Computer and Image Processing”, report “Modification of the Viola-Jones Approach
Russian Federation ”



Section “Image Processing Algorithm and Application”


What works we performed


This year we gave the following presentations:


  1. Fast Hamming distance computation on VLIW-architecture for 2D art recognition
  2. Methods of machine-readable zone post results processing
  3. Effective real-time training for a neural networks learning
  4. Document image recognition for mobile devices
  5. Viability of Viola-Jones method
  6. A method of projective transformations graph adjustment for planar object panorama stitching
  7. Seal of the Viola-Jones
  8. Hough image
  9. Spraying weighted image
  10. Dynamic programming approach to template-based OCR
  11. Recipe
  12. Image quality assessment
  13. Automatic cropping of images under projective transformation
  14. 2D art recognition in uncontrolled conditions using one-shot learning
  15. Differentiation of Common Image Processing Algorithms

As can be seen from the topics listed, we told not so much about documents as about general problems in the field of digital image recognition and effective ways to solve such problems.


Instead of conclusion


Perhaps, having read this publication, you, attentive reader, have a question: “Guys, you, of course, are great, but it’s not quite clear about what?” Therefore, we will now formulate our basic idea:



And in conclusion, another well-known fact: scientists tend to live long. Intense and constant mental work contributes to longevity, not allowing the body to age prematurely and go on a well-deserved rest. And this, by the way, is a rather strong motivational factor for a resident of modern Russia.


')

Source: https://habr.com/ru/post/431996/


All Articles