Software engineering photonics and color science

- A good photo is a combination of art and science
- Photo editing and color correction automation
- Color science theory combined with computer graphics techniques
- Spectral illumination data as a key for better machine learning
- Requirements for good photonics software or hardware applications

Transcript

1. Software engineering photonics and color science Petri Piirainen CTO & Co-Founder SoftColor Oy SOFTCOLOR SOFTWARE
2. BIO Petri Piirainen CTO & Co-Founder SoftColor Oy Born: 1981 Software business since 1996 2005: SoftColor Oy MSc computer science UEF alumni
3. Photonics applications Hardware / Software
4. Photonics applications Hardware / software
5. Physics Mathematics Electrical engineering Computer science Photonics applications
6. Physics Mathematics Electrical engineering Computer science Photonics application Arts • Image quality • Industrial design • User experience • User interaction • User interfaces
7. Software engineering photonics
8. Topics • SoftColor Oy • A good photo is a combination of art and science • Photo editing and color correction automation • Color science theory combined with computer graphics techniques • Spectral illumination data as a key for better machine learning • Requirements for good photonics applications
9. SoftColor Oy Products Company history Technology _SC
10. Faster, easier and affordable photo editing automation software
11. • Founded 2005 • Photo editing automation software for PCs, servers and cloud • Automata Pro • Automata Server • PhotoEQ • HQ: Joensuu, Finland SoftColor Oy
12. Customers: • Printing industry • Photographers • 360 photography • Newspapers • Ad agencies • Repro • Real estate • Car retail • Photo editors • Office tools • Developers • 85% USA + Canada + UK • Germany, Spain, Brazil Business
13. Our R&D
14. Our R&D Better and faster photo editing automation software
15. Computer vision + color science + computer graphics + DSP + machine learning SoftColor Engine
16. SoftColor Engine
17. Faster and easier photo editing automation
18. How does it work?
19. ORIGINAL AUTOMATICALLY ENHANCED
20. ORIGINAL AUTOMATICALLY ENHANCED
21. ORIGINAL AUTOMATICALLY ENHANCED
22. ORIGINAL AUTOMATICALLY ENHANCED
23. ORIGINAL AUTOMATICALLY ENHANCED
24. SoftColor & Photonics
26. Good photo / print = 80% art 20% science
27. SoftColor tone and color correction automation = SCIENCE ART +
28. Pixels Computer graphics Imaging algorithms Applied math Computer science Traditional image editing Photonics & color science
29. Pixels Computer graphics Imaging algorithms Applied math Computer science SoftColor Color and tone correction automation Photonics & color science
30. Processing colors not pixels
31. SoftColor tone and color correction automation SCIENCE ART Layer based processing Statistical analysis Metadata analysis Machine vision Machine learning Color grading for mass photo processing Selective color adjustments No workflow limitations
32. Processing colors not pixels More accurate automatic correction Batch color grading for photos Easier and accurate customization
33. SoftColor Engine
34. Layer based processing Original image Rich dynamics enhancer Luminosity enhancer White balance Natural color temperature Exposure and contrast Color grading
35. Layer based processing Rich dynamics enhancer Luminosity enhancer White balance Natural color temperature Exposure and contrast Color grading Enhanced image
36. Color and tone correction automation = Spectral illumination estimation technology
37. Automatic color and tone correction Original image Spectral illumination estimation from RGB image Metadata analysis EXIF & camera data Machine learning for estimated data = Automatic correction parameters
38. Color and tone correction automation Automatic correction parameters + User’s correction parameters = Enhanced image
39. Same parameters for human and computer Spectral illumination estimation from RGB image Metadata analysis EXIF & camera data Machine learning for estimated data Automatic correction parameters
40. The solution is possible by mixing spectral illumination data with computer graphics techniques Faster and more accurate automatic color and tone correction Same parameters for human and computer
41. Requirements for good photonics software or hardware applications
42. Software is the glue for photonics applications
43. Optical engineering Software engineering Electrical engineering User experience engineering
44. Optical engineering Software engineering Electrical engineering User experience engineering Battery and CPU limitations
45. Summary
46. - A good photo is a combination of art and science - Software is the glue for photonics applications - Photonics is a mixture of science, engineering, and arts
47. Thank you! Slides available at softcolorsoftware.com/blog petri@softcolorsoftware.com @softcolor_petri