The FoodSmartphone project has been reaching its half-time milestone. I have experienced a very steep learning curve during the first-half time period. And we have just spent the once-a-year Thanksgiving Day. Therefore, in the very beginning of this blog, I really want to appreciate people who came across and offered help to me. If it was not them, I would not have been able to travel this far. Standing at this point, I also feel grateful for all the opportunities of doing and presenting research offered to me. If it was not these opportunities and the amazing people who made them possible, there would have been no way for me to be in the FoodSmartphone project fulfilling my willingness of making the world a better place.
With all these supporting, we now have made one step forward – a step in the development of signal processing for the smartphone devices that we are working on. I would like to focus this blog on introducing this signal processing technique: What is it and Why is it important? Without further ado, let’s get into it..
One way to perform portable smartphone-based food inspection is coupling optical analysis with biochemical assays. That is, the light is used as the media for bio-sensing. There is a big portion of individual projects within the FoodSmartphone project focusing on developing optical bio-sensing techniques, including but not limited to colorimetry and Surface Plasmon Resonance (SPR). One of my major jobs in the project is to develop signal processing software to analyse these optical data and predict result from them. These optical data are mostly collected by using smartphone cameras. In the beginning of my research, I have noticed that images taken by different smartphones will appear in varied colour tones. This is true even with different individual products of a same smartphone model. The images above taken by different smartphone models demonstrate this effect – some of the photos show a reddish colour tone while some others showing blueish. Similarly, the flashlight on every smartphone exhibits its unique spectral characteristics – some of them are in warm and some others in cold colours. The smartphone flashlight is often used as the illumination source for smartphone optical analysis. The fact that the characteristics of smartphone camera and flashlight vary a lot will greatly and negatively affect the analysing of optical data since the colorimetry and analysis of spectrum are eventually achieved by measuring the colour values of images. Therefore, a technique to correct any effect due to variation in illumination source and smartphone camera is needed in order to perform accurate optical analysis. After the literature review, I have noticed that existing methods achieve such colour corrections by placing a standard colour reference chart in every image to be colour corrected. However, standard colour reference charts are hard and expensive to get. Thus, these methods stay unpractical for normal end-users.
To solve these issues, we have proposed a theoretical model that corrects the colour of images without the need of standard colour references. I am super excited that this work has been recently accepted by ISSPIT 2018 which will be held in Louisville, Kentucky, USA this December. A presentation of this work was also invited in the conference. It will be a great opportunity to demonstrate the vision of the FoodSmartphone project to the scientific community. Also, the research outcome from this work can potentially and tremendously improve the prediction accuracy of the optical analysis methods that we are working on in the FoodSmartphone project. I am going to implement this theoretical model in our FoodSmartphone signal processing software in the near future to further validate this technique using real applications in food analysis.
Okay, this blog stops here. I plan to explain you more technical insides of this work in my next blog. 2018 is reaching the end and Christmas is coming close. This means that a lot of traveling is ahead. At the end of this blog, I wish you safe traveling and a great time in this special season of the year!