Issues Log

09-08-2019: Delay in preparing and submitting a paper (Deliverable D2.2.)

03-09-2019: Problems on deploying the software on a server. We will only use it on the local machine.

20-06-2019: The budget amount for open access paper publication and the amount for subcontracting services not sufficient. Budget revised.

25-05-2019: Saturation and light of the specific image may result in altered colors (for example deep
saturated red may look like a light red to pink). We need a method that is resilient to the light conditions of the image. Possible Solutions: (1) Normalizations. (2) Reduce color on the image (transform the picture using a standardized color pallet (Windows 20‐color pallet or another pallet). Solution: Since the main objective of the project is to examine how effective is the concept of selecting an optimal safety color, we decided not to solve this issue. It can be addressed in a possible future commercial version.

20-05-2019: Various color quantization methods (mean, histogram, k-mean) fail in situations where the target has more than one main color. Solution: Advice to use windowing worked fine.

15-05-2019: The development of a background‐foreground color combination matrix ( Task 3.2) serves two purposes: (1) To have a table that can be used by practitioners to easily select the optimal safety color. (2) As a selection matrix for our software to chose the optimal safety color. We decided to cancel Purpose 2 since there are better ways to find contrasting colors (distance in the Color Space). For Purposed 1, backgrounds with two colors seem to be very important not to be exclusively addressed. Solution: To develop a more detail Background-Foreground matrix. Some delays in the planning are expected.

16-04-2019: Developing a mobile application just for WP5 is not cost-effective. Solution: Do great web application using Django framework (easier to maintain and possible to scale)

15-04-2019: As part of WP3, we are preparing a color combination scheme for optimal consciousness. The table is a mapping of 8 safety colors (….) to 16 background colors (representing the average of a domain in the color space). Two issues arise: (1) Which Color Space to use? (RGB, HSL, CIELAB, CIECAM). (2) How to divide the space of background color. Solution: For this project RGB is sufficient. Safety colors were reduced to 6 colors. For the background colors, a 20-color pallet is used (MS Windows 20-color pallet)

12-03-2019: Hiring a Software developer is not progressing fast. It is hard to find a developer with OpenCV experience and willing to work just for a few months. Solution: A second company employee will work on programming the user interface (on month August). Validation package was subcontracted

Scroll to Top