Paulina BugembeGreg DiamondFood Marketing and Image EnhancementECE 533Paulina BugembeGreg DiamondFood Marketing and Image Enhancement1Paulina BugembeGreg DiamondIntroductionWe began our project with a question: In marketing food products, whether it be via internet, print ads, etc., what sort of images sell your product?Certainly, health content, style of food, and a person’s personal taste have a lot to do with how we choose what to eat, but when it comes down to it, it all boils down to what “looks” tasty. How do we control and take advantage of what looks tasty to the human eye?After researching via the Internet, various food photography and marketing websites, we found that a common theme among food images is sharp images with the use of bright vivid colors with high contrast. From these characteristics, we derived a set of image processing techniques that could be used to change a perhaps not-so-appealing image to one that is appealing in the food market.These techniques include:- Saturation adjustment: to enhance washed out images- Adjustment of color: Varying the levels of red, green, and blue can provide for a pleasing effect for foods that are of the three primary colors.- High Pass Filtering: For sharpening edges in the images to provide a “clearer” picture.- Contrast Stretching: Changing the contrast level seems to be a pleasing enhancement to the human visual system.Using the following approach, we hope to show that these image enhancing processes canbe used to enhance images of food to make them more appealing to customers.Food Marketing and Image Enhancement2Paulina BugembeGreg DiamondApproachWe will implement the four image enhancement techniques shown above, and collect them into a small Image Enhancement software package. After creating the package, we will use it to apply various techniques and combinations of techniques on several different images of food.We will create an internet survey using the enhanced images to determine which enhancement technique achieved the desired effect. We will ask people to choose between two images, which image is the more appealing of the two: the original, or the enhanced image. We will run this experiment using several different images of different types of food to compare and contrast which techniques work better for making different types of food appealing to consumers.From the data obtained from this internet survey, we will try to decipher whether or not the techniques we chose made images more appealing to consumers.Food Marketing and Image Enhancement3Paulina BugembeGreg DiamondWork PerformedWe decided we were going to create a simple software package that included the four image enhancement techniques above, and a simple GUI. All code was implemented in C++.In Figure 1 below is a snapshot of the GUI for our software.Figure 1: Image Enhancement Software GUISaturation AdjustmentImages with low saturation appear to have “barely any color”, or are said to be “washed out.” This is because saturation of an image describes how much color is added to an image. An image with no saturation would be a monochromatic (grayscale) image. In our project, we altered the saturation value of an image by first converting its RGB color space coordinate values into HSI (Hue, Saturation, and Intensity) color space coordinate values. Then, we programmed three sliders into our GUI system, Hue, Saturation and Intensity. Although we didn’t change Hue or Intensity in our test images Food Marketing and Image Enhancement4Paulina BugembeGreg Diamondfor the project, since we had already converted to HSI coordinates, we could trivially create sliders for those two characteristics as well. Once the HSI values were changed according to the values that the user selected using the sliders, we converted the HSI coordinates back to RGB coordinates, and then displayed the image.Color AdjustmentSome images have a certain predominant primary color in them, and to emphasize that color could make the image of that food more appealing to the consumer. To create this functionality, we created three Sliders on the GUI in which users could vary the value for each of the Red, Green, and Blue color channels separately.High Pass FilteringSome images can look “blurry” or “fuzzy,” which can lessen its appeal to the consumer. To sharpen an image, we implemented a Laplacian high pass filter into our software package. The filter that we used is:-1 -1 -1-1 8 -1-1 -1 -1A Laplacian filter enhances the edges in an image, making the image appear “sharper.”Contrast StretchingFor contrast stretching, we implemented the Power Law transform on each of the Red, Green, and Blue color channels. The transformation function is:s = T(r) = c*rγThis transform is a pixel-based enhancement method that compresses the darker range of the color and enhances the lighter range making for a more even spread across the range of the color. Our GUI’s slider changed the gamma value for the Power Law transformation function.Food Marketing and Image Enhancement5Paulina BugembeGreg DiamondEnhancement of the Test ImagesAfter creating the image enhancement software package, we then used several test images to test the functionality of our software. The images we used are in the figures below.Figure 2: Chips n Salsa ImageFigure 3: Strawberry Cake ImageFood Marketing and Image Enhancement6Paulina BugembeGreg DiamondFigure 4: Tuxedo StrawberriesFigure 5: Fruits and VeggiesFood Marketing and Image Enhancement7Paulina BugembeGreg DiamondFigure 6: CheeseburgerFigure 7: Fried CalamariWe looked at each of the above images, and using the knowledge of what is considered tolook “good” in the field of food imagery, we chose which image would benefit most fromFood Marketing and Image Enhancement8Paulina BugembeGreg Diamondwhich enhancement, or combination of enhancements. In the table below, the name of the image, and which enhancements were applied to it are listed:Image Name EnhancementChips and Salsa Color Adjustment, Saturation, Combination of the TwoStrawberry Cake High Pass Filter, Saturation, Combination of the TwoTuxedoStrawberriesContrast Stretching, Saturation, Combination of the TwoFruits andVeggiesHigh Pass Filter, Saturation, Combination of the TwoCheeseburger Contrast Stretching, Saturation, Combination of the TwoCalamari Color Adjustment, Hue, Combination of the TwoTable 1: Image and EnhancementsWe implemented the website using Dreamweaver, HTML,
View Full Document
Unlocking...