The objective of this study was to analyze a set of images and classify the ones that showed any sign of damage on the plastic wrap. To accomplish this, I developed a C++ program capable of locating the file (text file) that contained every existing image address inside a folder and would then locate the matching image for every image address in the text file. There were image addresses for which the program could not find a matching image due to the absence of an image with a matching identifier. Furthermore, each image was classified using a basic pattern recognition program to increase productivity. The bar graphs below show the results of the classification process.
The classification process consisted of labeling each image with a code, the meaning of each code is described in the following list:
One of my first professional and enriching experiences was with the department of Soil and Crop Sciences at Texas A&M University at College Station. I collaborated on the research project “Heritability and quantitative trait loci of composition and structural characteristics in sorghum grain”. For this project I developed an algorithm in MATLAB to quickly estimate the area of the structural properties of sorghum kernels. The research team was using ImageJ, Thresholds and polygon tool, to manually measure the outer endosperm, embryo and inner endosperm. With this last method one person was only able to analyze five images per day, with the app one person was able to analyze 800 images per day. The results generated using this app were then used as supporting information in the study.